Abstracts - The 5th Congress of Polish Statistics

Objective

Development of a new algorithm for PCGLASSO—a scale-invariant estimator of the high-dimensional partial correlation matrix—and analysis of the properties of both PCGLASSO and the proposed algorithm.

Methods

Theoretical analysis of the properties of PCGLASSO. Derivation of the asymptotic distribution and the condition for the asymptotic recovery of the true structure of the related graphical model. Comparative computer simulations and real data analysis.

Results

The method can recover the hub structures much more efficiently than Graphical LASSO. The proposed algorithm is much more efficient than earlier solutions for PCGLASSO.

Conclusions

The proposed method is effective for uncovering the true structure of network dependencies and has broad applicability, for example in genetics and finance.

Keywords

precision matrix,graphical models,high-dimensional data,sparsity

Objective

The study aims to develop and compare advanced regression models, including novel variants of random forests, for functional data. We examine whether methods combining functional data analysis with random forests better capture complex relationships in the data than traditional statistical approaches.

Methods

Classical linear regression, functional regression, principal component analysis, and random forests were applied. Two new models were introduced: the Functional Random Forest (FRF), which incorporates functional analysis at the sample level, and the Full Functional Random Forest (FFRF), which performs functional analysis at the tree split level.

Results

The new functional models, FRF and FFRF, more effectively capture complex relationships in functional data than standard random forests and linear regression. Functional models also achieve good performance, particularly in specific applications.

Conclusions

Functional random forests (FRF and FFRF) offer an innovative approach to modeling functional data by combining the strength of random forests with functional data analysis. The study introduces new methods to statistics that can enhance data analysis in natural and technical sciences, inspiring further research on hybrid models.

Keywords

random forest, functional data analysis, principal components analysis

Objective

In this paper the hypothesis about the covariance matrix of a multivariate t distribution having the structure belonging to commutative quadratic subspace is considered

Methods

The maximum likelihood estimators of unknown parameters are used to determine three classical tests: likelihood ratio, Rao score, and Wald test.

Results

The forms of considered test statistics under the multivariate t distribution with known or unknown degrees of freedom, and under the multivariate distribution, are presented. The asymptotic distribution of all three test statistics for increasing sample size is also presented.

Conclusions

Simulation study evaluates the performance of considered tests showing that the Rao score test outperforms remaining two tests with respect to the speed of convergence to the asymptotic distribution. Moreover, it can be applied even in high-dimensional settings, when the number of unknown covariance parameters exceeds the sample size.

Keywords

Multivariate t distribution, Covariance structure, Rao score test, Likelihood ratio test, Wald test

Objective

This paper presents an empirical exploration of selected approaches to modelling the impact of the interaction of community well-being and individual (family/household) well-being based on official statistical data. The ultimate aim is to identify the best-fitting model for a specific analytical task, accounting for the limitations of the available data as input to a multi-source analytical database (instead of the desired but unattainable hierarchical (nested) data structure).

Methods

The assessment of interaction effects becomes crucial both from a methodological perspective and from an evidence-based (local) development policy perspective. It is assumed here that explaining such a complex problem requires considering both temporal and spatial aspects of inter-level dynamics together with relevant interdependent variables. In general, all the approaches discussed here constitute an example of a bottom-up, data-driven analytical strategy that differs from the model-driven strategy typically used in small-area estimation.

Results

The first part discusses measurement issues, with particular emphasis on the time-space approach (with reference to functional data, as formally the best) in comparison to approaches suggested by the type of problem or aspect of dependence of individual (subjective) well-being on the well-being of the local community, characterized by a multidimensional local deprivation index. The adequacy of the compared models to individual problem-data combinations is illustrated by calculations on official statistics data.

Conclusions

In conclusion, taking into account the limitations related to the use of existing data, an eclectic approach is considered preferable, pointing to the model that is appropriate to a given problem-and-data set. In the ideal case (under access to hierarchical geostatistical data), multilevel modeling of the interactions of well-being measures in a spatial context is considered the best.

Keywords

community-and-individual well-being, functional data, multilevel modeling, model choosing.

Objective

The analysis of imprecise experimental data requires the use of methods that take into account two types of uncertainty: randomness and imprecision. While classical mathematical statistics effectively handles randomness, the modeling of imprecision falls within the scope of fuzzy set theory. Unfortunately, the development of new hybrid analytical tools through the straightforward adaptation of well-known statistical techniques is rarely successful. Therefore, the presence of imprecise data necessitates the creation of new tools for statistical inference.

Methods

We propose a generalization of the Mann-Whitney and Jonckheere-Terpstra tests for imprecise data. Due to the lack of a linear order on fuzzy numbers, pairwise comparison of observations - central to the aforementioned tests - has been replaced by a dominance measure adopted from Liu’s uncertainty theory. Furthermore, the difficulties associated with determining the distribution of test statistics, which are essential for calculating p-values, have been overcome through the construction of permutation tests. The proposed tests have undergone extensive numerical studies.

Results

The application of Liu’s dominance measure in nonparametric permutation tests for fuzzy data—based on pairwise comparisons of observations—has proven satisfactory, especially when compared to results obtained using other tests described in the literature.

Conclusions

The proposed tests may be useful in problems involving the comparison of distributions of two or more populations, particularly in situations where the available data are imprecise (especially when the so-called human factor plays a significant role in their generation).

Keywords

location tests, imprecise data, fuzzy numbers, permutation tests, Liu`s domination measure

Objective

We propose a new goodness-of-fit test for the composite null hypothesis of exponentiality.

Methods

The new solution is based on an appropriately weighted empirical process with an estimated parameter. The test statistic is an Anderson-Darling type solution, while the weight is selected in such a manner that the random variables being the values of the introduced empirical process are asymptotically N(0, 1) distributed under the null model.

Results

We provide the tools that allow one to distinguish between different types of deviations from exponentiality. Such an approach sheds new light on a possible classification of popular alternatives considered in the literature. The asymptotic distribution of the test statistic under the null model is derived. It is also shown that the related test is universally consistent. An extensive simulation study and real data examples demonstrate the behaviour of the new test in practice.

Conclusions

We propose an omnibus test, new directional measures of discrepancy from exponentiality, and investigate both forgotten and popular procedures. The formulated theoretical outcomes provide the asymptotic distribution of the test statistic under the null model, as well as consistency under the alternative. The conducted simulation study clearly demonstrates that the new solution competes well with the best tests, whereas the directional measures of disparity from exponentiality provide insight as to the possible sources of deviations from the null model being a strong support for inference.

Keywords

Composite hypothesis, consistency, exponential scale family, goodness-of-fit test, invariant test, omnibus test.

Objective

Important

Methods

Research

Results

Important

Conclusions

Important

Keywords

sensitive questions, crosswise model, triangular model, confidence interval

Objective

The aim of the study is to identify areas in Poland that are similar in terms of the studied feature, i.e. the level of alder pollen concentration.

Methods

Functional data analysis methods were used to identify regions that exhibit similarity in terms of the level of observed values. Statistical techniques such as principal component analysis and cluster analysis were used to develop the experimental results. This approach allowed for effective capturing of complex relationships and grouping of objects with a similar variability profile.

Results

The obtained results of the analyses enabled the identification and delimitation of areas characterized by similar levels of the analyzed pollen concentration. The statistical methods used allowed the capture of spatial patterns of the concentration distribution, which in turn enabled the indication of regions with homogeneous aero-allergen properties.

Conclusions

The obtained results indicate significant advantages resulting from the use of a functional approach in the analysis of complex biological problems. The use of functional data analysis methods allows for taking into account the full structure of the temporal-spatial variability of the studied phenomena, which translates into increased precision in the identification of patterns.

Keywords

functional data: pollen concentration

Objective

The main goal of the paper is to present a new, two-version flexible distribution that is a modification of the Weibull lifetime model. An innovative idea is to replace the Weibull shape parameter with a shape function. Additional goals are to propose an estimation method and an information criterion. An extensive literature review is performed on 160 generalizations of the Weibull distribution, considering modalities and shapes of the hazard rate function. Cumulative failure functions of the bimodal models with a bathtub hazard rate function are given.

Methods

Properties of the proposed distribution are presented. Unknown parameters of the model are estimated using four method, namely: the maximum likelihood, least-squares, weighted least-squares and least absolute values.

Results

This article presents a three- and four-parameter flexible modified Weibull lifetime model called the Weibull distribution with a linear shape function. An innovative idea is to replace the Weibull shape parameter with a shape function. An estimation method based on theoretical and empirical reliability functions is proposed and the information criterion is defined. An extensive literature review was performed, taking into account the modalities and shapes of the risk rate function. To check the suitability and flexibility, the new lifetime models are validated with three real datasets and com

Conclusions

This paper shows that even a three-parameter distribution can compete with lifetime model with twice as many parameters in data modeling. The proposed Relative Reliability Criterion certainly helps to compare models when parameters are estimated by any method, but it does not consider the number of parameters (see AIC) and the sample size (see BIC, HQIC).

Keywords

Lifetime models, Weibull distribution, bimodal distribution, bathtub hazard rate function.

Objective

The paper addresses the analysis of the differentiation of local labour markets in the Wielkopolskie province at the county level, utilizing advanced dimensionality reduction methods based on principal component analysis: PCA, KPCA, and FPCA. An important aspect of the study is the comparison of the applied models based on the obtained empirical results.

Methods

The study employed three methods: classical principal component analysis (PCA), kernel principal component analysis (KPCA), and functional principal component analysis (FPCA). Cluster analysis using Ward’s method was performed based on the first two principal components. The data were sourced from the Local Data Bank and concern 12 variables describing the labour market in the counties of the Wielkopolskie province for the years 2004-2022.

Results

The clustering results were similar across all methods; however, a significant increase in explained variance was observed exclusively in the case of functional principal component analysis (FPCA) compared to classical PCA. FPCA models data as a function over time, preserving the continuity of trajectories and enabling the capture of key dynamic phenomena such as trends and seasonal patterns characteristic of labour markets.

Conclusions

The proposed methodology is flexible and can be successfully applied in other research areas, such as studies on poverty, socio-economic development or disability. It is particularly valuable in analyses where both feature variability and temporal dynamics play a crucial role in understanding complex socio-economic phenomena.

Keywords

functional data, FPCA, principal component analysis, labour market

Objective

The paper will present a graphical model describing the co-occurrence of allergic diseases.

Methods

We present this model in two versions: a generative model that reflects causal relationships and an approximation of the generative model using an ill-specified model that is computationally more efficient and easy to interpret. In both versions of our model, we consider typical symptoms of allergic diseases and additional covariates.

Results

The proposed model was evaluated using data from an epidemiological study on the occurrence of allergic diseases in Poland (www.ecap.pl). The bootstrap and jackknife techniques were used to assess the stability of the proposed model.

Conclusions

Our results show that the ill-specified model is a good approximation of the generative model and helps to predict the prevalence of allergic diseases.

Keywords

Bayesian network, logistic regression

Objective

One of the most important aims in insurance is to estimate a premium, which depends on a risk variable (number and severity of claims). The problem of robustness of the collective and Bayes premiums under uncertainty of prior knowledge is considered. The inaccuracy of the prior knowledge concerns the disturbance of independence between variables describing the frequency and average value of claims. Traditionally these variables are independent, but in applications, it is not always the case. The aim is to calculate the influence of dependence and to find the optimal premiums.

Methods

Two classes of priors are presented: in the first one, the FGM copula is applied, in the second one, the dependence between two contaminated priors is presented. In both classes, priors have the form of a linear combination of known bivariate probability distributions. The net individual premium is estimated and the square error loss function, to calculate collective and Bayes premiums, is applied. The optimal premiums are calculated under the minimax criterion.

Results

The ranges of collective and Bayes premiums are calculated and prior and posterior regret gamma-minimax premiums, as the optimal premiums, are presented. The results are applied to a numerical example. Despite the very mild or small dependence, its influence on the premiums is quite significant, especially on the bonus-malus factor.

Conclusions

Under considered classes of priors the optimal robust statistical procedures (estimators and predictors) are calculated. The situation, where the dependence between the frequency and mean severity of claims does not have the influence on the value of the collective premium is presented. In the example, we see that the dependence can produce significantly different Bayes premiums compared to the case of independent variables, even if the correlation coefficient is near 0.

Keywords

classes of priors, FGM copula, ε-contamination, posterior regret Γ- minimax premium, square error loss, bonus-malus factor

Objective

Item count techniques (ICTs) are statistical methods of indirect questioning that are broadly used by applied researchers in surveys with sensitive questions. These techniques require the use of some control variable while the variable under study is not directly observable. The aim of the paper is to examine the robustness of various item count models against some assumption violations.

Methods

To analyze the problem of robustness of item count models we conduct a comprehensive Monte Carlo simulation study. We introduce different violations in data distribution and examine the consequences for the efficiency of the estimation of the sensitive population proportion.

Results

For small departures from the idealized theoretical distribution of the control variable, estimators obtained by ML formulas via the EM numerical algorithm are still either more efficient or equally efficient as MM estimators in all considered cases (despite the fact that MM estimators do not depend on the distribution of the control variable in item count models). For moderate departures from the theoretical distribution of the control variable, in most cases the results are similar but there are several exceptions.

Conclusions

The parametric approach broadly used in ICTs to address the latent variable has many advantages in terms of estimation. However, it also introduces some problems regarding theoretical assumptions about the distribution of the control variable. We have analyzed the robustness of various item count models against departures from these theoretical assumptions. The presented analysis is of special importance for further applications of item count models.

Keywords

sensitive questions, item count techniques, latent variable, robustness, EM algorithm

Objective

To compare institutions, such as hospitals, schools, universities, social-care institutions and local authorities on a set of criteria. To discuss various formats in which the outputs of such analyses are presented and how well they fit the original purpose or remit of the comparison.

Methods

Historical review of the strengths and limitations of methods, such as simple comparisons of means and proportions, regression models, random coefficient models and methods based on causal inference, applying the potential outcomes framework.

Results

Two statistical learnings will be emphasised: the failure of hypothesis testing as a vehicle for making intelligent decisions and superior intellectual coherence of causal inference, including the potential outcomes framework, over modelling.

Conclusions

Methods that combine causal inference and decision theory enable a close match of the statistical method to the purpose of a comparison. They are more transparent and less dependent on assumptions that are peripheral to the purpose.

Keywords

Administrative database: causal inference: league table: random coefficient models: small-area estimation.

Objective

Estimating of population and domain means based on model-design approaches is considered. Population elements randomly belong to domains. A joint distribution of the variable under study and an auxiliary variable is assumed.

Methods

Data are observed in a sample selected from a fixed population. The partition of the sample elements into domains of the population is known. Outside of the sample, values of the auxiliary variable are known but their partition among the domains is not known. The domain means are estimated based on the likelihood function of the data observed in the sample and outside of it.

Results

The maximum likelihood estimation method provides the regression-type and the ratio-type estimators of domain means of the variable under study. They are dependent on the posterior probabilities that observations of the auxiliary variable belong to the particular domains. The weighted means of the domain averages estimators are used to estimation of the population mean. The accuracy of the evaluated estimators and the ordinary estimator are compared using a simulation analysis.

Conclusions

The results of this paper could be useful in economic, demographic and sociological surveys.

Keywords

domain means in a population, maximum likelihood estimation, regression estimator, ratio estimator, posterior probability.

Objective

The aim of the study was to evaluate the properties of complex estimators of population parameters that are functions of population mean values under the Pathak sampling scheme.

Methods

This objective was achieved by applying Taylor linearization, involving an unbiased estimator for the covariance between population mean estimators. Additionally, computer simulation methods were employed.

Results

Formulas for the approximate bias and approximate mean squared error of complex population mean estimators were derived. Approximate variance estimators were proposed.

Conclusions

Proposed estimators for approximate bias and approximate mean squared error of the complex population mean estimators respectively converge to the bias and mean squared error values obtained through simulation.

Keywords

complex estimators, Pathak`s scheme, unbiased estimator

Objective

The main objective of the paper is to present the economic activity of population 50 and over in the context of the decision to retire in Poland compared to other European Union countries. The study presents the differences in the age at which retirement begins and the share of people receiving benefits. The analysis includes the main reason for ceasing work upon retirement, as well as the main reason for continuing to work upon retirement or returning to the labour market.

Methods

In the study, multidimensional statistical and comparative analysis methods were used. Using demographic analysis methods, changes in life expectancy and age structures were also presented, as well as the consequences for potential labour force resources and the dependancy ratio. The study used Eurostat data, in particular from the 'Pension and labour market participation' module conducted in 2023 as part of the European Labour Force Survey. In addition, data from Statistics Poland and OECD were used.

Results

The results of the study indicate that despite the increase in economic activity observed in recent years, the retirement age in Poland is among the lowest in Europe. The participation of people receiving benefits in the labour market has increased – 11.6% of people aged 50-74 who receive a pension continue to work. On the other hand, over 75% of retirees aged 60-64 remain outside the labour market. This applies to 94% of women and 33% of men. After retiring at the age of 65-69, the share of men outside the labour market increases to over 97%.

Conclusions

The demographic conditions of shrinking labour resources and changes in the age structure indicate the need to develop incentives for longer economic activity.

Keywords

pension, economic activity of population 50+

Objective

The aim of the presentation is to determine the degree of willingness of people in pre-retirement age to work after obtaining pension rights. It also identifies the demographic and social characteristics that currently characterise those (un)interested in extending their working lives.

Methods

The data sources are the individual results of a questionnaire survey of the population at pre-retirement age (10 years to retirement age) conducted on a nationwide representative sample of the Polish population. Statistical analysis was conducted using selected multivariate methods.

Results

A low willingness to continue employment or start new job after retirement age was observed - it was considering by only one in five people (usually assuming working for 3-5 years, in their current position, full-time). On the other hand, half of people are of the opposite opinion, 2 / 3 declaring retirement at 60(K) / 65(M). Among the factors differentiating these opinions are work ability, place of residence (including distance to the poviat / voivodeship capital), level of education and type of job, while gender or being in the protective age are insignificant.

Conclusions

The research makes it possible to estimate the number of people who, if declarations of continued work are transformed into real behaviour, will be working beyond retirement age in 5-10 years.

Keywords

people in pre-retirement age, economic activity, retirement

Objective

Balancing professional and family roles remains significantly more difficult for women than for men, as women are more often burdened with caregiving and domestic duties. Although some progress has been made, change is slow. The COVID-19 pandemic forced many to reorganize their work and family lives. This article aims to assess the extent and durability of changes in women’s professional and domestic responsibilities in Poland, as well as public attitudes toward women’s position in the labor market.

Methods

The study utilizes data from the Labor Force Survey (BAEL, 2010–2021) describing labor supply on both extensive and intensive margins: preliminary results from the Time Use Surveys (2004, 2013, 2024): the European Social Survey (ESS, 2010, 2021, 2023) on attitudes toward gender and family roles: and the European Working Conditions Telephone Survey (EWCTS, 2021) on the perceived ability to balance work and private life and time burdens from family duties. The article also outlines the key findings and limitations of these datasets.

Results

Despite the observed changes, women continue to be disproportionately burdened by responsibilities associated with their family roles. Some of the changes identified during the pandemic period proved to be temporary.

Conclusions

Lasting transformations in the patterns of balancing professional and familial roles require the support of appropriate family and labor market policies.

Keywords

survey, labour supply, work-life balance

Objective

Resilience can be seen as an individual`s ability to adapt and develop despite adversity and age-related challenges, but it can also be viewed as a relational social process. It can therefore be assumed that an individual`s resilience can be a factor determining the exchange of intra- and intergenerational support. The main goal of the presentation is to present the results of analyses regarding receiving and providing care (for adults or grandchildren) taking into account markers of resilience in people aged 50+.

Methods

Data from the ninth wave of the Survey of Health, Ageing and Retirement in Europe (SHARE) conducted in 2021 / 2022 were used for the analyses. A logistic regression model was estimated for dependent variables describing the fact of providing or receiving care. To determine the level of resilience of individuals, latent class modeling was used for variables describing psychological well-being, health status and subjective financial situation. Basic socio-demographic and economic characteristics of individuals were included as control variables in the models.

Results

The results showed that receiving care may be associated with characteristics describing lower resilience of individuals. Moreover, care is provided to individuals who belong to latent classes of lower resilience, and at the same time, lower resilience leads to a lower probability of providing care to both adults and grandchildren. Higher social capital, measured for example by social networks, increases the chances of both receiving and providing support.

Conclusions

To conclude, to build a resilient society in the future, individual resilience and the factors that influence it (education, employment) need to be strengthened, which will translate into greater independence of individuals and exchanges of support between them.

Keywords

Keywords: resilience, care, SHARE, older people

Objective

This study offers a comparative analysis of fertility transition in Romania and Moldova – two Eastern European countries with shared historical roots but divergent socio-economic paths since the 1990s. The research aims to identify features of the fertility transition, assess the extent of birth postponement, estimate fertility recuperation in completed cohorts. There is a lack of comprehensive comparative research on fertility in Romania and Moldova, as both countries have often been excluded from major Eastern European studies due to concerns about data quality.

Methods

Fertility change is approached through the concepts of postponement (decline at younger ages) and recuperation (increase at older ages). The analysis uses vital statistics on births by mother’s age and birth order from the National Institute of Statistics of Romania and the National Bureau of Statistics of Moldova (1971-2023). Tomáš Frejka’s four-phase fertility transition model guides the interpretation, while the decomposition method distinguishes between fertility among younger (15–28) and older (29–49) women. The Gamma Fertility Curve was applied to estimate changes in fertility.

Results

Results show a sharp fertility decline in both countries in the late 1980s and early 1990s. Romania entered the transition in 1988-2001 (TFR minimum: 1.22), while Moldova followed, reaching its lowest TFR (1.44) in 2002. Romania has since progressed to more advanced stages of recuperation at older ages, whereas Moldova remains in an intermediate phase, with slower recovery.

Conclusions

These countries reflect distinct post-socialist fertility transition patterns, with common trends of delayed childbearing and reduced fertility, yet shaped by different historical and institutional contexts. The study highlights the need for differentiated demographic policies and contributes to broader fertility transition research in Central and Eastern Europe, addressing the underrepresentation of Romania and Moldova in comparative studies.

Keywords

fertility transition, fertility postponement and recuperation, Frejka model, Romania, Moldova

Objective

The purpose of this paper is to analyze seasonal variations in the monthly number of deaths in Poland and selected European countries in the first two decades of the 21st century using the TRAMO / SEATS method. The analyses cover European countries from different geographic regions, characterized by significant differences in the health status and age structure of the population.

Methods

The study used the TRAMO / SEATS method in the JDEMETRA+ software, which is recommended and dedicated by Eurostat for time-series analysis and deseasonalization. The long time series available made possible to examine differences and similarities between countries, as well as to assess the stability of seasonality patterns and their changes over time. Data were obtained from the Eurostat database and Statistics Norway.

Results

The results provide a broad picture of seasonality of deaths in the countries studied. The higher amplitude of the number of deaths and greater susceptibility of the population to seasonal factors in the south than in the north of Europe were confirmed. Decreasing role of seasonality of the death rate, encountered in many studies, was not confirmed. We note differences: the lowest amplitude of fluctuations in the west is in the Netherlands, the highest in Spain and the United Kingdom. In the south, Italy (large and increasing amplitude) and Greece (lowest amplitude, summer extreme) stand out.

Conclusions

The use of the time series decomposition method for analyzing demographic data to extract and fully characterize the pattern of seasonality of death is a novel approach used in this work. The issue of seasonality of death and its measurement addressed in this work is part of not only the growing interest in the scientific community in this issue, but also has great significance for health policy and the identification of necessary strategies in the situation of recurrent epidemics and the challenges posed by environmental factors and cyclically changing weather conditions.

Keywords

seasonality of death, TRAMO / SEATS decomposition, SARIMA models

Objective

The aim of this presentation is to introduce the U-smile Likelihood Evaluation method (U-smile rLR), which allows for assessing the predictive usefulness of a variable in binary classification without the need to define a decision threshold. The method separately evaluates prediction quality for events and non-events, offering more interpretable insights than classical approaches.

Methods

The U-smile method is based on the decomposition of the relative Likelihood Ratio (rLR) and its visualization using U-smile plots. It enables the identification of variables with asymmetric predictive impact across classes and assessment of their predictive strength. The method was applied to the Heart Disease dataset, with random (non-informative) variables used as a reference.

Results

The analysis showed that variables with high rLR values produce distinct 'smiles' on U-smile plots, which are easy to interpret visually. The plots allowed us to determine which group (event or non-event) contributed to the statistical significance of the Likelihood Ratio test. The findings were consistent with those based on ROC curves. The method proved robust against non-informative random variables and predictive asymmetry.

Conclusions

U-smile is an effective, threshold-free method for evaluating predictive variables that complements ROC and calibration plots. It provides a graphical and statistical assessment of predictive strength and class asymmetry, in line with Explainable ML principles. Its intuitive visual format facilitates result communication, even to non-technical stakeholders.

Keywords

binary classification, rLR, U-smile, predictive model evaluation, model interpretability

Objective

In the recent years, there has been a dynamic, steadily deepening decline in the population of Poland. This process is shaped primarily by natural increase, which has remained negative since 2013. A record high natural decline was observed in the past 2024, when 156.7 thousand more people died than were born (a decline of 41 people per 10 thousand population). The aim of this paper is to present the most important demographic factors influencing unfavorable trends in mortality and birth rates.

Methods

The source of the data are the results of demographic surveys conducted by the Statistics Poland – collected by the Statistical Office in Olsztyn. The characteristics of population reproduction processes were characterized by the following coefficients: demographic dynamics, natural increase and fertility. The observed trends in the intensity of births and deaths were referred primarily to the transformations in the population structure by age. Various forms of graphic presentation were used to visualize the analysis results, such as: cartograms, age pyramids and charts.

Results

The dynamic decline in the number of births (apart from the procreative behaviors of the population) is influenced by the decreasing number of women of reproductive age, especially those aged 20-35. In the case of deaths, the greatest impact on their intensity is the increasing percentage of people aged 65 and over in the total population (the ongoing process of ageing of population). Additionally, in 2020 and 2021, the covid-19 pandemic had a particular impact on the rapid increase in the number of deaths. Analyses also showed clear spatial differentiation in terms of vital statistics.

Conclusions

The observed demographic processes may significantly affect all areas of socio-economic life in Poland. The demographic forecast indicates that the current trends regarding natural decrease and ageing of population will deepen.

Keywords

natural increase, depopulation, vital statistics, ageing of population

Objective

The aim of this presentation is to introduce capture-recapture methods for population size estimation (dual-system estimation, multiple-system estimation) that account for record linkage errors and their implementation in the {uncounted} package in R. The package is currently being developed at the Population Research Methodology Center in the Statistical Office in Poznań.

Methods

The presented tool includes classical estimators as well as estimators that incorporate record linkage error correction (one-way, symmetric, and asymmetric (Di Consiglio * Tuoto, 2018: Zult et al., 2025)).

Results

The package offers the capability for variance estimation and construction of confidence intervals to assess the quality of the estimates.

Conclusions

As part of the presentation, we will showcase the package in the context of administrative data, with special emphasis on modeling record linkage errors. Examples with simulated data will illustrate the package`s capabilities in various scenarios.

Keywords

capture-recapture, population size estimation, administrative registers, record linkage errors

Objective

Development and release of a universal, error-tolerant tool (package) for record blocking that minimizes the number of comparisons in entity resolution procedures without sacrificing matching accuracy, especially under conditions of missing identifiers and low data quality.

Methods

Approximate Nearest Neighbours, Graph indexing, record blocking

Results

Development of a package that provides high recall while achieving a high reduction of pairs required for comparison.

Conclusions

The package is a tool suitable for use in integrating data from administrative registers for the purpose of population size estimation. A scientific article describing the proposed method is currently under review in the journal SoftwareX.

Keywords

record linkage, deduplication, approximate nearest neighbours, official statistics

Objective

Identification of further residence plans of Ukrainian citizens residing in the Łódź and Silesian voivodeships, i.e. in voivodeships with a diversified economic context.

Methods

A logistic regression model was used to analyze the further residence plans of Ukrainians in the Łódź and Silesian regions. The explanatory variables included, among others, those illustrating the situation of immigrants on the labor market, knowledge of the Polish language, their satisfaction with selected areas of life (financial, housing situation) and the type of support received. The main source of data is a survey conducted among Ukrainians living in the Łódź and Silesian voivodeships in the fall of 2024. The tool used was an online survey - in total, the survey covered 680 Ukrainians.

Results

The results obtained indicate significant differentiation in terms of further residence plans, both among migrants who arrived before and after the war, as well as between the analysed voivodeships. The results of the analysis indicate, among other things, that knowledge of Polish increases the probability of declaring a desire to stay in Poland, especially in the case of war refugees. People who declared satisfaction with their financial situation expressed over 1.6 times greater willingness to stay in the Łódź region, and 1.4 times greater in the Silesia region.

Conclusions

The conducted study makes it possible to estimate the number of Ukrainians who are willing to stay permanently in selected regions and the factors that increase the chance of staying in the region they currently live in.

Keywords

Ukrainians, labor market, Lodz Voivodeship, Silesian Voivodeship

Objective

The study presents the demographic processes of the Łódź and Świętokrzyskie voivodeships in 2002-2022. The following research goals were set: analysis of changes in the population of counties and municipalities in 2002-2022: identification of areas with annual population loss in the period under study: identification of factors of population change in these areas (typology): analysis of the population ageing (including classification of municipalities in terms of similarity in changes in the age structure of the population).

Methods

In the study selected descriptive statistics tools were used, as well as selected cluster analysis methods. The classification of counties according to population change determinants was also carried out. The source of statistical data were the publications of the Central Statistical Office for counties and municipalities of the Łódź and Świętokrzyskie voivodeships in 2002–2022.

Results

The obtained results illustrate the demographic problems of the Łódź and Świętokrzyskie voivodeships, characterized by a relatively high intensity of the depopulation process (in many municipalities as a result of both natural and migration decrease) and a high level of demographic aging. Particular attention was paid to counties and municipalities characterized by annual population loss in the period 2002-2022.

Conclusions

The study provides a synthetic picture of the demographic situation in the central macroregion over two decades. The obtained results identify areas of population growth and areas of the greatest, often long-term demographic decline. The main factor of population changes in the counties and municipalities of the central macroregion is natural decrease. Migration-related population loss is of particular importance in the depopulation process in the case of unattractive areas, especially those located on the outskirts of voivodeships.

Keywords

depopulation, ageing process, determinants of population changes, central macroregion,

Objective

The aim of the paper is to measure the extent of extreme poverty, deprivation, and food poverty in Poland. The central research question is whether the so-called original OECD equivalence scale, used by Statistics Poland (GUS) to estimate extreme poverty and deprivation, is consistent with the scale effects implied by the construction of the subsistence minimum and the social minimum for different household types. In addition, the paper proposes using food expenditure thresholds – adopted in the measurement of extreme poverty and material deprivation – as boundaries for defining food poverty.

Methods

The research focuses on the concepts of the subsistence minimum and the social minimum. To estimate equivalence scales, one-parameter power scales were applied. The analysis is based on anonymised data from the 2023 Household Budget Survey conducted by Statistics Poland.

Results

The findings indicate that the equivalence scale used by Statistics Poland assumes a stronger scale effect than that derived from the subsistence minimum, which leads to differences in the estimated extent of extreme poverty. Indicators of food poverty were also estimated.

Conclusions

The results are of significant importance for improving poverty measurement methods and for informing social policy based on more precise poverty methods.

Keywords

extreme poverty, deprivation, food poverty, equivalence scales, poverty rate

Objective

The aim is to assess the reasons for the existing of pensions below the guaranteed minimum pension from the perspective of persons belonging to the group 'new poor retirees'.

Methods

The data source is the individual results of a questionnaire survey of the peri-retirement population (women aged 55-65 and men aged 60-70) conducted on a nationwide sample of individuals who, due to their short insurance tenure, are classified as current (NBE) or potential (PNBE) 'new poor retirees'. The study was carried out as part of the 'New poor retirees' project (IPiSS). Statistical analysis was carried out using selected multivariate methods.

Results

The number of 'new poor retireers', whose insurance tenure is lower than 20 (F) / 25 (M) years, is increasing year by year. In 2024, 433,100 such benefits were paid out, thus affecting nearly 10% of new-system pensioners, far more often women than men. Among the main reasons for being placed in the NBE group were: the difficulty of finding a job, caring for children and others, working illegally and lack of economic necessity to work, and among the long-term reasons were health problems, childcare, sharing responsibilities with a partner and working abroad.

Conclusions

A better understanding of the reasons for an insurance tenure that is too short to obtain a pension at least at the level of the minimum pension will allow the development of more relevant social policy solutions.

Keywords

economic activity, retirement, pension below the minimum level

Objective

The analysis of differences between poverty measures is mainly based on income levels or indicators of material deprivation. Their comparison has consequences, such as differences in assessing the scale of the problem. The research has two goals: methodological and practical. The methodological goal is to present a methodological concept for studying poverty as a barometer that will inform in advance about changes in the situation of the population. The practical goal is to apply this methodology to study poverty in territorial units in Poland.

Methods

Our research focuses on a multidimensional approach based on multi-criteria decision-making methods that consider various aspects of poverty, such as health, education, housing conditions, and access to services. We use data from the Local Data Bank of Statistics Poland, which allows us to obtain accurate and up-to-date information. We aim for the research to be repeatable, enabling the monitoring of changes over time and the adaptation of social policies to dynamically changing conditions. Additionally, we identify the time of exiting or deeper falling into poverty of territorial units.

Results

This analysis represents the essence and extension of research on the diagnosis of poverty and social exclusion in the Mazowieckie Voivodeship from 2012 to 2023. The findings emphasize the need for an integrated approach to identifying poverty, one that considers various indicators influencing poverty.

Conclusions

An integrated multidimensional approach to identifying poverty, which considers various indicators, is crucial for effectively monitoring and adapting social intervention strategies. The research emphasizes the need to consider different aspects of poverty in social policies, which allows for a more comprehensive understanding of the problem and better adaptation of interventions to the actual needs of impoverished individuals.

Keywords

poverty, objective measures, social policy, multidimensional approaches

Objective

The study aimed to diagnose disparities arising in the financial situation of households headed by a person with higher education against other household classes between 2010 and 2023. The analysis approached the financial situation as a multidimensional concept, including income levels, expenditure patterns, savings rates, and the burden of fixed expenses on household budgets. The research confirmed that the growing supply of higher-educated individuals in the labor market contributed to a dynamic decline in the financial returns to education, as reflected in households’ financial situation.

Methods

The study drew on data from the Household Budget Survey conducted by Statistics Poland. The analysis classified households by educational attainment, focusing on the group with higher education. To assess disparities in the financial situation from a multidimensional perspective, the study applied methods of relative taxonomy. For each year, it calculated relative synthetic indices for the higher education group against other groups.

Results

Households with higher education consistently demonstrated better financial situation than all other classes between 2010 and 2023. Still, this advantage declined significantly over time. Across all examined dimensions of financial situation, the study observed a narrowing of disparities between the higher education class and the remaining household classes – notably in terms of disposable income, where it was most dynamic.

Conclusions

The relationship between education and wages has been extensively studied in literature. Empirical research on the education premium focuses on individuals with higher education or specific occupational groups. This study offers a novel perspective by examining how the financial returns to education of the household head extend to the entire household. Moreover, it captures a wider range of outcomes by considering various dimensions of financial situation. By applying methods of relative taxonomy, the study was able to diagnose disparities in financial situation across education classes.

Keywords

financial returns to education, household financial situation, relative taxonomy

Objective

The aim of the study is to compare factors influencing saving and saving for retirement by young adults in Poland and the Czech Republic. Based on the behavioral life cycle hypothesis, it was assumed that behavioral factors exert a significant influence on young people`s attitudes toward saving. The variation in factors relates not only to differences between the countries studied but also between generations.

Methods

The study used survey data collected in December 2021 among young adults from generations Y and Z. A binary logit model and pairwise comparisons were used as a research tool.

Results

The study looked at behavioral, financial, demographic, and educational factors affecting general and retirement savings among young adults of late Generation Z and early Generation Y in Poland and the Czech Republic T The results reveal similarities and differences in savings behavior both across age groups and between the two countries. The greatest diversity was identified in behavioral factors: in Poland, the endowment effect and hedonic consumption influence retirement savings, while in the Czech Republic, personal financial responsibility is a key factor. Common factors influencing young

Conclusions

The findings point to the need for targeted education and retirement policy solutions.

Keywords

Savings, behavioral life cycle, retirement, young adults

Objective

People with disabilities constitute 16% of the world`s population, and this number is growing due to the aging of the population and the increasing incidence of non-communicable diseases. Disability is a multi-faceted problem - economic, social and health, having a profound impact on the quality of life of individuals and their economic stability. The aim of the conducted research is to analyze the economic situation of people with health-related activity limitations for selected European countries, focusing on income, financial difficulties and poverty in the surveyed households.

Methods

The research was carried out using data from EU-SILC 2023, which allows for a comparative assessment of six countries representing Central and Eastern Europe (Poland, Czech Republic, Hungary, Slovakia - the Visegrad Group) and Western Europe (Germany, France - which together with Poland form the Weimar Triangle). The theoretical Dagum distribution for equivalent income expressed in PPS was used to analyze income, and methods of analyzing the structure, income gap and poverty rate of the economic situation by disability level in the studied countries were also used.

Results

The results reveal a strong association between activity limitations and economic hardship, with severe limitations leading to greater difficulties in meeting basic needs and higher poverty rates. These patterns were most evident in Germany, France, Slovakia and Hungary. The Dagum model confirmed income disparities, with people without limitations receiving the highest incomes. Hungary and Germany had the largest income gap, while Slovakia had smaller disparities.

Conclusions

The results show that reducing economic disparities between able-bodied and disabled people requires a comprehensive approach that includes financial support, wider access to education, inclusive employment, reduced healthcare costs and appropriate anti-discrimination regulations and measures. Inclusive policies are key to reducing inequalities and improving the well-being of people with disabilities. The results of the analysis also point to the role of national policies in mitigating disparities due to disability.

Keywords

disability, economic situation, international comparisons

Objective

This paper addresses the growing relevance of unpaid household and care work in the context of increasing demand for care services due to demographic and labour market changes.

Methods

To valuing unpaid work the replacement cost method (input method) using market wage rates was applied. This method assumes that the input is the time spent on specific domestic tasks and activities. Time-use data were drawn from the 2023 Time Use Survey, and hourly wage rates by occupation were used to estimate the monetary value of household work. Selected macroeconomic aggregates from national accounts relating to the household sector were then incorporated into the valuation of non-market household production. A Household Production Satellite Account for Poland 2023 was compiled.

Results

The valuation revealed that unpaid household work accounted for approximately 90% of non-market household production in 2023 (compared to 78% in 2013). The gross value added of non-market household production represented approximately 40.4% in relation to Poland’s GDP in 2023. Given its significant scale, this form of production should not be excluded from official public statistics. It can be presented through a satellite account of household production as a complementary data source alongside national accounts, offering insights into the real contribution of households to the economy.

Conclusions

The substantial size of unpaid work and non-market household production confirms the contribution of households to both the care economy and the informal economy. It also allows for meaningful comparison with GDP. The Satellite Account of Household Production fills an important gap in the system of social and economic statistics by estimating the non-market contribution of households. Alongside national accounts, it serves as a valuable source of information on time transfers and household resources.

Keywords

care work, informal economy, unpaid work, gender equality, household production satellite account, time use survey

Objective

The aim of this presentation is to highlight the potential for incorporating insights from international projects focused on older adults into official statistics.

Methods

In this context, the presentation will showcase the results of a survey conducted as part of the international City*Co project in two Polish cities – Kraków and Wrocław – on a sample of 804 individuals aged 65 and over. The study was carried out using face-to-face interviews conducted in respondents’ place of living (42.7% of the sample) and telephone interviews (57.3% of the sample). In line with the adopted methodology, a Polish-language tool was developed and validated to measure the age-friendliness of cities and communities (AFCC), based on the concept proposed by the WHO and Dikken et al. (2020).

Results

The reliability indicators of the tool were as follows: Cronbach’s Alpha – 0.875, McDonald’s Omega – 0.876. Based on the AFCC scale assessments and using hierarchical cluster analysis, three segments of older adults were identified, and the differences between groups of residents in both cities were characterized.

Conclusions

The developed tool and the obtained results can be used to design public policies that are more socially acceptable and more effectively address the needs of older urban residents. The tool is already recommended for implementation in other countries, which may contribute to the inclusion of these issues in future social research agendas. An additional value of the project was the testing of different ways of reaching older respondents – both face-to-face and telephone interviews. The results indicate that for this age group, in-home face-to-face interviews are more effective.

Keywords

ageing policy, evaluation, social indicators, survey research

Objective

The ageing population of nurses in Poland is becoming a serious problem, and the generation gap is widening. In addition, the availability of staff and predictions of a shortage of staff in the 20–40 age group are important issues. The aim of the research is to find out the views of nurses aged 18-44 on job satisfaction and the factors that can influence it, and to identify a group of stressors that contribute to leaving the profession.

Methods

The study involved a group of 404 nurses from the Mazowieckie and Świętokrzyskie voivodships and used an auditory questionnaire. Workplace stress factors were classified according to the Levi and Frankenhaueser system. Non-parametric analysis methods were used for results determined on ordinal scales.

Results

Statistically significant differences were found between nurses` age and their ratings of social relations at work and their degree of stress in terms of physical working conditions, time pressure and coping with new technology. A significant relationship was also found between nurses` age and their decision to undertake retraining. A group of stressors was identified as the reason for the decision to leave the nursing profession.

Conclusions

Nurses aged 18–34 experience workplace stress more acutely and are more likely to consider a career change, especially if nursing was their first career choice. Providing appropriate working conditions, support from supervisors and help with managing emotions is crucial for retaining nurses in the profession.

Keywords

nurses, job satisfaction, occupational stress, survey

Objective

The aim is to present the work in progress at the UNECE forum and to promote the discussed topic of work quality (from the employee`s point of view).

Methods

The study is based on an expert method of labor market statisticians.

Results

The effect of the research is a discussion on the developed list of indicators measuring the quality of work.

Conclusions

A list of indicators for measuring the quality of work has been developed.

Keywords

quality of employment measurement

Objective

Monthly pay (wage) from the main job is one of the mandatory variables in the European Labour Force Survey (EU-LFS, in Poland: BAEL). When answering the question referring to it, respondents often tend to provide rounded values. This causes a distortion of the distribution of observed values, which takes on discrete features and has consequences for the way data is processed and analysed. In particular, it may disturb the process of nonresponse imputation. The aim of the paper is to present these problems and discuss the applied solutions.

Methods

The paper presents the scale of the phenomenon and the implications for the method of imputation of missing data, as well as the consequences for the analytical use of data, e.g. when determining quantiles of the distribution. Due to the connection with the analysed phenomenon, the technical elements of the imputation procedure are also discussed. The presented examples and empirical analyses concern the Polish LFS data for the years 2021-2024.

Results

The results show a significant scale of the occurrence of rounded values. It is shown how this can result in a bias of the imputation results and what measures were taken to remedy this. The consequences regarding quantile groups are presented. The issue of disturbing rounded values in order to make the distribution closer to continuous is discussed.

Conclusions

The significant scale of rounded values requires a deeper consideration of the interpretation of point data on wages obtained in the survey. Conclusions regarding the distortion of rounded values, both in the context of imputation and outside this context, in a broader sense, are referred to the Eurostat guidelines for the EU-LFS.

Keywords

labour force survey, imputation, deciles, discrete distribution, continuous distribution

Objective

Studies on resilience in developmental psychology, demography, sociology and economics show that current adaptability in the face of shocks depends on factors experienced across life stages. This study aims to assess how the optimality of life trajectories (calculated for ages 20–50) affects quality-of-life changes during the pandemic among people aged 50+, compared to the influence of other socio-demographic characteristics.

Methods

Firstly, the Needleman–Wunsch algorithm was used to perform the sequence analysis. Each sequence contained yearly elements representing one of the 33 socio-economic situations from SHARE Wave 7. Secondly, the results of sequence analysis were treated as a predictor of changes in life quality during the pandemic (change in the CASP index between the period of 2015–2017 and 2022), using multilevel logit models and partially proportional odds models with clustered standard errors.

Results

Belonging to the quintile group characterised with the most financially optimal life paths was associated with a 7.3-percentage-point lower risk of quality-of-life decline during the pandemic, compared to the least optimal group. The effect of a one-standard-deviation change in life trajectory alignment was 4-5 times lower than a one-standard-deviation change in loneliness or depression levels (measured with a short R-UCLA scale and the EURO-D scale). This effect was similar to the one-standard-deviation change in education level.

Conclusions

The effect of favourable life trajectories aligns with the notion of resilience and relates to the literature on developmental psychology (Waters and Sroufe, 1983). The Needleman-Wunsch algorithm has been used in social sciences since the 1980s and more recently in research based on the SHARE data. This study stands out by the thorough computation of substitution costs between sequence elements, with a substitution cost matrix based on dissimilarities between respondent groups calculated using a modified Gower metric.

Keywords

SHARE, resilience, sequence analysis, COVID-19, multilevel analysis

Objective

The presentation shows the results of the work on the Building and Dwelling Database.

Methods

Data from the 2021 Population and Housing Census, data from the NOBC system (Address identification system for streets, properties, buildings and dwellings), survey B-07 – reports on residential buildings and residential units in non-residential buildings (put into use) and Attrition were used to create the database.

Results

The use of data from administrative sources and non-statistical distributed systems for statistical purposes will contribute to broadening the scope of studies and analyses carried out in the field of housing and municipal management.

Conclusions

The Building and Housing Database will comply with the EU requirements for European statistics on population and housing (ESOP) and will form the basis for building and housing censuses. Combining data from different sources will allow the creation of new, richer datasets, more fully describing the phenomena under study and enabling analysis of relationships between information that is available in separate sources.

Keywords

Housing stocks

Objective

The aim of this study is to present the role of reintegration units in preventing social exclusion in Poland between 2015 and 2023. These units implement social and professional reintegration services targeted at long-term unemployed individuals, persons with disabilities, people leaving penitentiary institutions, and the homeless. Their operation was strengthened by the Act on social economy of 2022, which defined their status within the social economy sector and assigned them a key role in implementing the objectives of the National Programme for the Development of Social Economy.

Methods

W analizie wykorzystano dane pochodzące wyłącznie ze źródeł administracyjnych: systemów informacyjnych urzędów wojewódzkich, Ministerstwa Rodziny, Pracy i Polityki Społecznej oraz Państwowego Funduszu Rehabilitacji Osób Niepełnosprawnych. Wyniki W 2023 r. działało 1246 jednostek reintegracji społeczno-zawodowej – o 17 mniej niż rok wcześniej, ale o 57 więcej niż w 2015 r. W ciągu 2023 r. z usług reintegracji społeczno-zawodowej skorzystało blisko 49,0 tys. osób zagrożonych wykluczeniem społecznym, tj. o 2,1% więcej niż w 2022 r. W latach 2015-2023 ich liczba wzrosła o 21,6%. Działalność prowa

Results

In 2023, 1,246 units providing social and professional reintegration services were active—17 fewer than in 2022, but 57 more than in 2015. Nearly 49,000 individuals at risk of social exclusion benefited from reintegration services in 2023, representing a 2.1% increase compared to 2022 and a 21.6% increase compared to 2015. Most of the entities focused on activating people with disabilities—733 occupational therapy workshops and 141 vocational activity establishments.

Conclusions

Reintegration units play an important role in preventing social exclusion by supporting individuals at risk of marginalisation in returning to the labour market and active social life. Ongoing monitoring of reintegration units is essential to assess the impact of their services on the economic independence of individuals at risk of social exclusion, which is crucial for effective social policy and sustainable social development.

Keywords

Reintegration units, social exclusion, social and professional reintegration services, social economy

Objective

The objective of this presentation is to present the findings of an experimental study focused on identifying the occupations held by individuals actively engaged in the national economy (excluding the unemployed).

Methods

The study utilized data drawn from administrative registers and statistical surveys conducted by official statistics. A key innovative element was the approach that considered all employment positions held by an individual, regardless of the type of contract. This enabled a more comprehensive capture of actual professional activity. The research focused on the occupation actually performed, rather than the one formally acquired through education or training. This required precise identification of the nature of work performed in a specific place and time.

Results

The resulting data open up new opportunities for labor market analysis by allowing for a detailed representation of employment structures by actual occupation, broken down by demographic, territorial, and sectoral dimensions.

Conclusions

This study holds considerable relevance for both socio-economic policy and vocational and educational planning, supporting evidence-based decision-making (data-driven policy). It opens new perspectives for monitoring structural changes in the labor market and for improving the alignment of the education system with the real needs of the economy.

Keywords

labour market, occupation held, administrative data

Objective

The presentation aims to show how modern tools of the official statistics system – based on administrative data – are revolutionising the way wages and salaries are analysed and enabling a more dynamic and detailed view of the labour market.

Methods

The Survey on the Distribution of Wages and Salaries in the National Economy is entirely based on administrative sources – mainly registers of the Social Insurance Institution (ZUS).

Results

The results of the survey provide a unique and regularly updated picture of wage differentiation in the national economy. The data are published at a monthly frequency, with an approximately five-month lag from the reference period, which ensures that the wage situation can even be monitored on an ongoing basis. The first data published in September 2024 concern January, February and March 2024. The results of the survey are published in the form of a news release and a set of tables. The data are presented with breakdowns both by characteristics of a person earning a wage or salary (sex, age,

Conclusions

The use of administrative data in statistical surveys is not only in line with global trends in modern statistics systems, but is becoming the foundation for a more responsive and accurate description of socio-economic phenomena. The range of published data and information is being expanded. Data on wages and salaries published every month allows the phenomenon to be analysed using the average wage / salary and the median wage / salary with unprecedented detail. These statistics are becoming a dynamic tool for monitoring changes and making decisions in near real time.

Keywords

wages and salaries, distribution of wages and salaries, median wage / salary, average wage / salary

Objective

The aim is to investigate if there is an effective yet automatic procedure for detecting downsizing and upsizing? An attempt is made to construct such a procedure in the R environment. Moreover, the aim is to determine the scale of hidden inflation for different food and non-food product groups and the sensitivity of different price index formulas to the phenomenon of over- and under-sizing (disproportionate to its price).

Methods

The study uses IT toosl: R environment and PriceIndices package. The research method is an empirical study using actual scanner data. The data source is a retail chain operating in Poland (cooperating with Statistics Poland).

Results

Both groups of products that do not generate hidden inflation (e.g. bread) and groups of products that are particularly vulnerable to downsizing and upsizing (e.g. yoghurt) are indicated. The results suggest that price index formulas (weighted and unweighted) have different vulnerabilities to hidden inflation (being more or less sensitive to downsizing and upsizing).

Conclusions

Multilateral indices have been shown to be much less sensitive to the phenomena under consideration than bilateral indices. In particular, chained indices (e.g. the Jevons chained index) appear to be particularly susceptible to the phenomenon of hidden inflation. The results in this regard are pioneering in the literature.

Keywords

scanner data, product downsizing, shrinkflation, product upsizing

Objective

The aim of the study is to identify the strength of the statistical base's impact on the value of the current CPI inflation index

Methods

The study is based on indicator analysis, time series analysis, R package, Statistics Poland database, NBP dataset.

Results

The reasearch results indicate a decomposition of the inflation rate to assess the statistical base effect.

Conclusions

The statistical base affects the value of inflation rates and should be taken into account when interpreting inflation rates for monetary policy purposes.

Keywords

inflation, CPI, statistical base, monetary policy

Objective

The aim of the paper is to examine the relationship between inflation, measured by the year-on-year CPI index, and the value of assets of manufacturing enterprises in the public sector in Poland. The analysis seeks to determine whether an increase in consumer prices affects the size of these firms’ assets and in what direction this relationship occurs.

Methods

The study applies vector autoregressive (VAR) models and Granger causality tests using quarterly data from the 2014–2024 period. The data sources include quarterly reports on financial assets and liabilities (RF-01) published by Statistics Poland (GUS) and data from the National Bank of Poland (NBP).

Results

The VAR model shows that changes in CPI inflation precede changes in the assets of manufacturing enterprises. The Granger test confirms a statistically significant direction of influence from inflation to asset dynamics.

Conclusions

The results suggest that CPI inflation can serve as an effective predictor of changes in the assets of manufacturing enterprises in the public sector. These findings have implications for economic policy and confirm the usefulness of VAR models and Granger causality tests in sectoral research, contributing to the empirical literature on the public sector in Poland.

Keywords

inflation, enterprise assets, VAR models, Granger causality test

Objective

The paper presents methodological aspects of BTS, its development as well as use of data for additional analyses and applications. Additional elements include the impact of war in Ukraine on the situation of enterprises in Poland and rapid structural diagnosis of macroeconomic processes at micro level (Mic-Mac). Moreover, the results of additional work using data from the survey will also be presented (output gap, employment gap, but also business cycle indicator or analysis of business cycles for the main economies of the world and the EU, included in the annual Macroeconomic Report).

Methods

Additional sets of questions are included in the standard BTS. Initially, they concerned the COVID-19 pandemic, then this issue was replaced by questions concerning the war in Ukraine. Currently, the additional module includes questions on basic macroeconomic processes: investments, labour market, price developments, which provide – rare in official statistics – behavioural aspects of companies` decision-making. With the use of data from the survey as well as those in the field of business climate analyses, additional elements used in other studies and analyses are also developed.

Results

Data gathered from these BTS extensions provide – together with the results of a standard part of the survey – information on the current assessments of the impact of the war in Ukraine on business activities as well as up-to-date key economic issues. The first results were presented within a few days after the survey was carried out in very detailed cross-sections.

Conclusions

Due to its scale, subject matter and level of detail, the survey along with its developmental elements are a source of quick information on the consequences of emergency situations, selected aspects of business activity or key macroeconomic processes from the company`s perspective. The survey results allow the provision of additional analytical information on important phenomena including the cyclical position of the economy, output gap or labour hoarding. Further developments are planned, which will gradually transform the survey into a comprehensive tool for multidimensional economic diagnostics.

Keywords

business tendency, factors limiting activity, output gap

Objective

The aim of the study is to examine the application of non-random data sources in business statistics.

Methods

The study uses the simulation method on the basis of a superpopulation model built from the frame and dataset of a real business survey.

Results

The study shows a comparison of the accuracy of the used estimators with the unbiased Horvitz-Thompson estimator.

Conclusions

Estimators using mass estimation are feasible in terms of the acceptable levels of accuracy compared to the representative method.

Keywords

mass imputation

Objective

The aim of the study is to present an analysis of the activities carried out to develop PKD 2025 and the promotion of the classification in order to encourage domestic users to independently re-register their PKD activity codes in official registers, including REGON, KRS and CEiDG.

Methods

The study shows the statistics and analysis of the impact of the wide-ranging national consultations on the PCAs in the new classification. The changes introduced, in particular regarding the divisions of the current groupings, should be duly justified (for example, turnover or a significant estimated market share).

Results

As a result of meetings organized by Statistics Poland and promotion of changes to the PKD 2007 in the years 2022-2024, 71 institutions, chambers, business associations, government administration units and the UAS applied. The largest group of entities that submitted proposals for changes to the classification were institutes, chambers and associations of entrepreneurs. As part of a broad consultation on the amendment to the PKD 2007, 173 proposals for changes to the classification were collected. In addition, as part of the legislative process, further changes and new c

Conclusions

Classification of PKD activities is a classification with a wide application, used, among others, for reporting, registration and economic analysis purposes. For these reasons, it is important that it reflects the changing reality, i.e. it takes into account new technologies, new activities emerging on the market, new categories of service activities.

Keywords

PKD 2025, PKD amendment, new PKD, classifications

Objective

• Practical use of the τ-Argus program on aggregated data.
• The rules of Statistical Disclosure Control in DG-1 Survey.
• The importance of statistical suppression in data protection.

Methods

Statistical disclosure control techniques can be defined as set of methods to reduce the risk of disclosing information on individuals, businesses or other organisations. SDC methods minimise the risk of disclosure to an acceptable level while releasing as much information as possible.
Statistical disclosure control primary and secondary suppression with the program τ-Argus.
We use these methods in the DG-1 survey: 'Assessment of the current economic activity of enterprises' (completed on the DG-1 form, 'Report on economic activity').
This is one of the main sources of information on the current economic situation of entities in Poland. The data obtained reflect the economic and social situation in the country and each of its provinces.
Form DG-1 is a monthly survey which includes legal entities, organisational units without legal personality, and natural persons with 10 or more employees whose activities have been classified according to the 2007 Polish Classification of Economic Activities (PKD) as falling within the scope of selected sections.

Results

Preparation of files needed for statistical disclosure control with the τ-Argus program.
Practical presentation of the steps statistical disclosure control primary and secondary suppression in τ-Argus program for single data files and related tables.

Conclusions

Since January 2016, statistical disclosure control in the DG-1 survey has been determined based on the value of the cumulative turnover variable (On). This means that the occurrence of dominance in the aggregate is determined by the variable On. The τ-Argus programme is used to determine which aggregates are covered by primary and secondary suppression.
We also use the τ-Argus program to analyze statistical suppression in other tables and surveys.
The τ-Argus program simplifies statistical disclosure control, especially secondary suppression.

Keywords

short-term statistics, DG-1 survey, statistical disclosure control, τ-Argus

Objective

This study aims to assess the financial effects of internationalization in the food industry by comparing exporters and non-exporters (using ROA, ROE and ROS), and by analyzing the impact of internationalization intensity (foreign revenue share) on these metrics. Hypotheses are divided into two sets: basic export status comparison and continuous level of internationalization.

Methods

The study applies fixed-effects panel regression, complemented by Hausman, Breusch-Pagan and robust standard error corrections (Huber-White). The dataset includes over 9,800 firm-year observations from 1998 to 2023. Dependent variables are log-transformed ROA, ROE and ROS, and independent variables include internationalization level, revenue, total assets and market capitalization.

Results

No significant differences in ROA, ROE or ROS were found between exporters and non-exporters. However, a higher level of internationalization significantly reduced ROA and ROS, with ROE remaining statistically unaffected. For ROS, a nonlinear, escalating negative effect was detected.

Conclusions

Internationalization does not automatically improve profitability in the food industry. Negative effects on ROA and ROS may reflect expansion costs, operational complexity and delayed returns. This study challenges common assumptions and contributes robust empirical evidence to the literature on firm internationalization.

Keywords

internationalization, financial performance, food industry, panel data, ROA, ROE, ROS

Objective

The aim of the study is to construct a macroeconomic model describing the impact of armaments on the dynamics of basic macroeconomic variables in the economy.

Methods

The model uses elements of the theory of differential equations and numerical methods,

Results

The model allows for simulations of military production growth paths mimicking the arms race process, for example during the Cold War between the USA and the USSR or between North and South Korea.

Conclusions

The model can be extended as a multi-equation econometric model, and the estimates of its parameters can describe actually functioning economies.

Keywords

Solow model, arms race, differential equations, numerical methods

Objective

This study investigates the determinants and distributional characteristics of production capacity utilization (CU) in the Polish economy, with a focus on its right-censored nature at 100%. Recognizing the central role of capacity utilization as both a business cycle indicator and an input in estimating production functions, the paper aims to construct a more comprehensive, sectoral disaggregated measure beyond the commonly used manufacturing proxy.

Methods

We use firm-level data from Business Tendency Surveys (BTS) and annual enterprise activity reports covering 2008–2023. Given the censored nature of the reported capacity utilization at 100%, we apply a Tobit panel data model to appropriately estimate the latent full capacity utilization levels. The analysis incorporates sectoral controls, output levels, investment outlays and energy-material expenditures to assess both firm-level and sector-level determinants.

Results

The results reveal strong sectoral heterogeneity in capacity utilization, with firms in more competitive markets and service sector reporting significantly higher utilization levels. The econometric evidence confirms that capacity utilization is strongly influenced by firm output, input use, and investment activity. The observed distributions display a substantial mass at the 100% threshold, supporting the case for modelling CU as a right-censored variable.

Conclusions

By accounting for right-censoring and sectoral heterogeneity, the proposed approach delivers a more accurate and informative measure of capacity utilization. This enhances our understanding of cyclical dynamics and provides improved inputs for productivity analysis and policy formulation. Identifying the determinants of capacity utilisation and the sources of sectoral and structural heterogeneity enables better alignment of economic policy instruments with actual enterprise conditions.

Keywords

capacity utilization, Business Tendency Surveys, panel-data Tobit model, sectoral heterogeneity, censored data

Objective

The aim of the paper is to present how the initiative of performing KLEMS productivity accounting by Polish voivodships was achieved, and to demonstrate its viability and usefulness. This includes the explanation of the basic methodology, the presentation of the necessary methodological extensions, the demonstration of the effectiveness of the applied methodology in the form of published results, and some examples of specific applications of final data.

Methods

The basic methodology relies on the idea of economic growth decomposition in the form of KLEMS growth accounting (otherwise called KLEMS productivity accounting) adapted to Polish data availability. A trans log production function is being used in calculations not only at the aggregate level but also at industry level. The special achievement of Polish KLEMS initiative is to develop a methodology for decompositions at regional level, i.e. at industry level and by voivodship at the same time. Data for the research have been provided by Statistics Poland’s units.

Results

The results of the calculations have been published on Statistics Poland’s site. They allow to study economic growth at regional level (by voivodship) and by industry. Labour, capital and multifactor productivity contributions (and their sub-contributions) to growth can be observed at the level of these aggregations, allowing for ample economic analyses of which the possibilities are yet to be explored.

Conclusions

The numerous possible applications include the possibility to study and compare not only economic growth and the contributions to it at the aggregate but also at regional level thus allowing to observe the effectiveness of regional economic policy at different NACE activity aggregations. The specific use can consist of studying specialisations, including smart specialisations of different regions. A specific industry example will be provided as a demonstration.

Keywords

KLEMS, productivity, economic growth, growth decomposition

Objective

Over the past decade, the influx of immigrants has been a major factor in the Polish economy. Since the Russian military invasion of Ukraine on February 24, 2022, a large wave of Ukrainian refugees has arrived in Poland. This article aims to estimate the contribution to labour supply and GDP growth in Poland of three groups of immigrants: Ukrainian refugees, previous economic migrants from that country and other immigrants.

Methods

The article applies the decomposition of Poland`s economic growth based on the production function. The approach utilised in the article also allows for the heterogeneity of capital and labour, including the differentiation of human capital and sectors and professions of Polish citizens and selected categories of immigrants. The study uses data from the LFS and surveys of immigrants conducted by the National Bank of Poland.

Results

The results of the decomposition of annual GDP growth dynamics by different types of production factors indicate that in the years 2021–2023, immigrants contributed on average 0.5 percentage point of annual GDP growth (18% of the total growth). In the 2013–2021 period, it was also 0.5 percentage point, although the higher overall dynamics meant that the share in the total growth was slightly lower (13% of the total growth).

Conclusions

The contribution of immigrants has helped maintain Poland`s high economic growth potential despite the gradual decline in total factor productivity and the negative impact of the ageing population of Poland.

Keywords

economic growth, TFP, work input, immigration, Ukraine, refugees

Objective

The aim of the study is to assess the level of uncertainty of different groups of respondents in relation to the answers given to the questions included in the survey.

Methods

The study used the Dempster-Shafer method, in which unlike Bayesian theory, instead of specific probabilities, the degree of conviction is assigned to the answers given. This allows for the assessment of the respondents' uncertainty in relation to a given question. The source of the data are the results of survey research conducted by DG ECFIN.

Results

The results of the conducted analysis allow for the identification of discrepancies in the degree of uncertainty of the answers given in periods of generally increased uncertainty.

Conclusions

The applied method allows for a better understanding of the cognitive structure among the surveyed entities. It can supplement the description of the results of the survey research conducted.

Keywords

uncertainty, survey research, probability

Objective

Agriculture is an important sector of the economy in both Poland and the European Union. It influences, among others, food security, jobs, rural development and the environment. For this reason, it is necessary to constantly monitor farmers` opinions regarding their activities. This is possible thanks to the business tendency in agriculture survey, conducted by Statistics Poland since 2012. The aim of the conducted analysis is to examine the impact of selected variables on the assessment of the situation of agricultural holdings in Poland.

Methods

Data obtained from the business tendency in agriculture survey (AK-R) completed by natural persons managing agricultural holdings were used in the research. They were analyzed using ordered response multinomial models.

Results

The obtained results will allow determining the degree of impact of individual variables on the assessment of the situation of farms.

Conclusions

They can be used to further verify the business tendency survey in the agricultural sector, as well as to support decision-makers in developing a strategy to support agricultural policy.

Keywords

business tendency survey, Polish agricultural holdings, ordered multinomial model

Objective

This study examines the influence of the Human Development Index (HDI) on Sustainable Development (SD) in EU countries by analysing the relationship between HDI and a new Sustainable Development Index (SDI), highlighting regional disparities and policy implications.

Methods

Using 18 Eurostat indicators from 2021, the SDI is constructed via the TOPSIS method. Bilinear ordering visualizes EU countries’ SDI and HDI scores. Spearman’s rank correlation analyses their relationship, with data normalized using zero unitarization and winsorization to handle outliers.

Results

The SDI ranges from 0.270 (Romania) to 0.709 (Denmark). HDI varies from 0.795 (Bulgaria) to 0.948 (Denmark). Northern / Western EU countries (e.g., Denmark, Sweden) score higher in both, while Eastern / Southern countries (e.g., Romania, Bulgaria) lag. Spearman’s correlation (r=0.852) shows a strong link.

Conclusions

The strong SDI-HDI correlation underscores their alignment. Northern / Western EU countries excel in both, while Eastern / Southern ones need targeted strategies. Policymakers can use these insights to balance human and sustainable development, enhancing EU-wide outcomes.

Keywords

Sustainable Development, Human Development, TOPSIS, Bilinear Ordering, EU

Objective

Many countries are implementing the Perpetual Inventory Method recommended by the OECD in a 2001 manual. In the Polish system of national accounts, the initial work just completed was used to test the fit of the gamma distribution density function to the retirement pattern for transportation means. In doing so, reference was made to the practices of the German statistical institute described in the Measuring Capital OECD ManualIn the further estimations conducted, the AN 1131 retirement order was developed for individual NACE divisions.

Methods

The paper will present the results of modeling the population exit rate function for commercial trucks. The primary base was a subset created from the national register Central Register of Vehicles and Drivers. The modeled exit functions - gamma and log-normal were tested for further application to PIM in Poland.

Results

The estimates data for investment outlays included time series for gross fixed capital formation and also list of price indices and service lives for different types of fixed assets in different kind of activities: calculations based on investment outlays and based on GFCF using PIM assumptions for different types of assets to obtain current values of asset stocks: calculations of consumption of fixed capital based on new estimates.

Conclusions

Paper includes information on the current approach applied in Polish national accounts to estimate the value of fixed assets, scope of methodological guidelines analysed that provides recommendations on PIM method, development of the assumptions for PIM in Polish national accounts, the results of experimental calculations and their impact on GNI.

Keywords

PIM, GFCF, depretiation function

Objective

The aim of the presentation will be to assess changes in the mobility of the population in Poland related to three types of movements: migration for permanent residence, commuting to work and commuting to schools. This type of mobility will be presented in systems related to the administrative division (voivodships) and selected typologies of municipalities.

Methods

The analysis will compare selected structural measures of graphs for the studies mentioned above, illustrating individual aspects of mobility. Selected R packages will be used and presented for loading data provided by official statistics, processing them and visualizing the obtained results.

Results

The obtained results will allow us to verify several hypotheses concerning population mobility. The research questions are as follows: how does population mobility change between successive editions of studies characterizing different types of movements? Are there regional differences in the dynamics of changes in structural measures of graphs characterizing population mobility? Are there differences in the dynamics of changes in structural measures of graphs characterizing population mobility in the systems of selected urban typologies?

Conclusions

The presented approach emphasizes the need for official statistics to increasingly broadly present statistical research results in a matrix format, which enables advanced analyses, including primarily those related to network analyses based on the graph theory. Examples of this type of information include permanent residence migration matrices, matrices showing the connections between the place of residence and the place of work, and matrices showing the connections between the place of residence of a student and the place of education.

Keywords

population mobility, graph theory, structural characteristics of graphs, typology of cities

Objective

One of the key problems in the development of world economies is the implementation of sustainable development goals. Some of the goals are focused on environmental factors mainly related to industrial operation and fossil fuel consumption. The concept of energy efficiency, which assumes the achievement of both carbon emission reductions and sustainable economic growth, can be used to evaluate the Sustainable Development Goals. The purpose of this study is to assess energy efficiency in European Union member states.

Methods

One approach to studying energy efficiency is to analyze energy intensity, defined as energy consumption per unit of output (usually gross domestic product). This approach is criticized for ignoring the structure of the economy and environmental conditions, thus assuming that economic output is the only determinant. Another approach is the DEA (Data Envelopment Analysis) method, a non-parametric linear programming method for measuring the efficiency of multiple decision-making units.

Results

The application of the DEA model assumed the following input variables: GDP, renewable energy, country area or population. The analysis also used the determinants of CO2 emissions: energy intensity, energy structure, carbon intensity and economic activity. As a result of the research, the potential of individual member states to reduce carbon emissions was determined and they were ranked by this criterion.

Conclusions

Ranking EU member countries by the criterion of energy efficiency made it possible to identify countries that are leaders in terms of reducing CO2 emissions and countries that need to intensify their efforts to improve the effectiveness of environmental policy.

Keywords

energy efficiency, CO2 emissions, Sustainable Development Goals, DEA, EU member states

Objective

The aim of this study is to assess the spatial differentiation of green areas in Polish regions using spatial data and official statistical data in relation to noise and pollution as factors influencing the quality of life of residents.

Methods

The analysis used statistical methods (correlation analysis, cluster analysis, median classification method, spatial analysis using Moran`s statistics). The results were additionally visualised on spatial maps using information from the geoportal (BDOT10k map) and OpenStreetMap.

Results

The availability of green spaces varies widely and not all countries and regions are equally equipped with this infrastructure. Some regions have better access to green spaces and clean air, while others struggle with higher levels of noise and pollution.

Conclusions

By producing maps showing the relationships between the availability of green spaces, air quality and noise levels in selected regions of Poland, it was possible to identify where green spaces have the greatest impact on achieving Sustainable Development Goals. The results provide recommendations for Poland and its regions, defining the scope and places where coherent action is needed to create better conditions for improving the quality of life.

Keywords

quality of life, green spaces, sustainable development, spatial analysis, regional development

Objective

The aim of the analysis is to assess the impact of European Union funds under the 2014–2020 cohesion policy on the level of innovativeness in Polish voivodships. The main research question is: does the intensity of EU support in the area of R&D translate into an increase in regional innovation capacity, or is the effectiveness of this policy primarily dependent on local absorption conditions?

Methods

The analysis is based on data from the Ministry of Funds and Regional Policy concerning the value and structure of funding agreements related to R&D activities, supplemented by innovation indicators from the European Commission’s Regional Innovation Scoreboard and statistical data from Statistics Poland (GUS) for the 16 Polish voivodships. The study employs correlation analysis methods to identify relationships between the level of support and innovation outcomes at the regional level. The results are also presented in the form of graphical visualizations.

Results

Between 2014 and 2020, the largest share of EU funding was allocated to enterprises, which had a particularly strong impact on regions with a higher capacity to generate innovation, such as Mazowieckie, Małopolskie, and Pomorskie. In the eastern regions, despite the use of EU funds, challenges remain in fully realizing their innovation potential, in part due to limited demand for innovation and lower levels of development. The effectiveness of innovation funds therefore depends not only on their scale but also on the ability of regional markets to absorb and implement project outcomes.

Conclusions

Impact of EU funds varies across regions. A key limitation remains the low capacity of many regions to implement innovations. The effectiveness of the funds does not depend on their amount, but on market conditions that determine the real possibility of translating projects into economic outcomes. The analysis contributes to the literature on regional development policy by highlighting the need to design policies based not only on the allocation of funds, but also on the actual capacity to use them effectively within the economy.

Keywords

regional innovativeness, EU funds, cohesion policy, regional development

Objective

In the face of escalating climate challenges and the urgent need for energy transformation, municipalities play a crucial role in implementing sustainable development policies. They undertake investment activities aimed at developing a low-carbon economy, the spatial distribution of which may exhibit significant spatial correlations. The objective of this study was to identify and assess the phenomenon of spatial autocorrelation in municipal investments related to the development of a low-carbon economy in the Greater Poland Voivodship, co-financed by European Union funds. The research aimed

Methods

The study was based on data obtained from the Local Data Bank of Statistics Poland (GUS) and databases of the Ministry of Funds and Regional Policy. To assess the spatial autocorrelation of municipal investments in the low-carbon economy, global and local measures of spatial autocorrelation were applied. The calculations were conducted in the R environment.

Results

The results revealed a spatial concentration of the intensity of municipal investments in the low-carbon economy across the Greater Poland Voivodship. The identified clusters of municipalities with high and low levels of implemented low-carbon projects confirm that functional urban areas may play a significant role as poles of intensification for this type of investment.

Conclusions

This study offers a new perspective in the literature on local sustainable development by integrating the analysis of low-carbon investments with a spatial approach. It demonstrates that the distribution of climate-oriented actions at the municipal level is not random but may be shaped by territorial dependencies and functional structures. Consequently, the research contributes to a deeper understanding of the spatial diffusion of the energy transition at the local level and enables the identification of territorial mechanisms that support or hinder the development of a low-carbon economy. The

Keywords

low-carbon economy, municipal investments, spatial autocorrelation, global Moran’s I, local Moran statistics, spatial analysis

Objective

The problem analyzed in the presented study concerns the differentiation of development paths of communes (gminas) – local development trajectories – within voivodships, in the period from accession to the EU (until 2016). And what was the impact of subsidies received by gminas on their level of development, characterized in terms of reducing the multidimensional index of local deprivation (MILD, for the years 2004-2016). Moreover, did this impact show any pro-convergence tendency? In particular, how do such issues look in the case of the Mazowieckie Voivodship.

Methods

The studied issues are analyzed with focus on the Mazowieckie Voivodship divided (after Eurostat) into 'capital' and 'region', using BDL data. The typology of development trajectories built on the basis of changes in the scope of MILD during these years allows, on the one hand, for insight into the factors of differentiation between them, and on the other, for the assessment of the policy of allocation of funds (subsidies) from the principles of spatial justice and efficiency standpoints (using Marginal Benefit Incidence Analysis) and their impact on spatial inequalities.

Results

As expected, the relatively most different trajectories ('growing development vs. 'declining') were observed between both parts of the voivodship (including urban-rural distinction). While subsidies generally go to the 'most needy' gminas (high on the MILD scale), and in accordance with MBIA, convergence consists in reducing inequalities among the least developed (levelling down) rather than 'up', with a tendency towards spatial clusters (according to I-Morano), e.g. in the southern part of the Mazovian province.

Conclusions

Apart from the limitations of the information contained in the Local Data Bank, it allows for the use of many analytical tools suggested in the literature for assessing development processes. Both at the stage of planned distribution of development funds (it is possible to estimate the objective demand for development, ex ante), and for evaluation of the 'spatial justice' and effectiveness of the scheme of their allocation (ex post).

Keywords

local development measurement, 'spatial justice', marginal benefit analysis (MBIA), spatial autocorrelation

Objective

The aim of the presentation is to show the results of an experimental study on the spatial analysis of the effects of the flood in Poland in September 2024 and the applicability of the developed method to monitor the effects of natural disasters.

Methods

Remote sensing data, geospatial data and administrative data were used for the study. The survey consisted of 3 stages. In the first, the extent of flood-affected areas was determined on the basis of satellite imagery from the Copernicus programme (Sentinel 2). In the second stage, address points were assigned to the delineated flood zone extents (source: National Population and Housing Census 2021), and then data on persons and buildings from the integrated administrative database ('List of Persons for Experimental Study') were assigned to the address points.

Results

Linking data from such a variety of sources made it possible to obtain, for the areas actually affected by the floods, information such as the population and its demographic characteristics, employed persons, economic entities, farms, buildings and their age, and others. For the purpose of performing spatial analyses, the data were aggregated to the municipality, census area and district. The talk will present the main conclusions of these analyses and detailed statistics for several of the most affected municipalities.

Conclusions

The developed data integration method is a complex process, in which the main goal at each stage is to ensure high data quality (including coherence and comparability). At the same time, it provides great opportunities to prepare, in a relatively quick way, spatial analyses of the extent and effects of floods, as well as other types of natural disasters.

Keywords

Earth Observation methods, remote sensing, geospatial data, administrative data, flood

Objective

In many local government financial supply systems, shares in state taxes are an important source of revenue. Although they do not have taxing power over them, they are counted among the fiscally efficient sources of revenue for local government units (LGUs). Therefore, it is reasonable to consider the evaluation of this source of revenu for territorial self-government units, particularly in light of the legal changes that took effect in Poland on 1 January 2025. This presentation aims to determine the level of diversity among Polish communes in terms of their income from PIT participation and

Methods

Descriptive statistical measures and graphical presentation methods will be employed to evaluate the level of diversification of the analysed source of revenue for territorial self-government units. Conversely, econometric modelling will be employed to identify the determinants of fiscal efficiency. The data used in the analysis comes from the Ministry of Finance's resources and official statistics (Local Data Bank). The Statistical Geoportal will be used to prepare the presentation.

Results

Although shares in the PIT are quite significantly differentiated spatially, they constitute a fiscally significant source of revenue for territorial self-government units. Small rural municipalities (predominantly agricultural) receive significantly less revenue from PIT than urban municipalities or cities with county rights. According to the simulation, the changes introduced on 01.01.2025 will lead to a reduction in these disparities, although some units will still require significant subsidies under the existing income equalisation schemes.

Conclusions

The results presented here can be used to evaluate the fiscal efficiency of the analysed revenue source of LGUs. They may also provide a basis for further in-depth analyses of the importance of this source of revenue in shaping the financial self-reliance of LGUs.

Keywords

tax sharing, LGU's own revenue, financial self-reliance, fiscal efficiency determinants of PIT share

Objective

The paper aims to develop and evaluate, using Polish data, a simple spatiotemporal method for disaggregating annual regional GDP series to quarterly frequency using a smoothing algorithm, and to compare its results with a multivariate regression-based approach that uses auxiliary variables.

Methods

The method assumes that the disaggregated series should exhibit minimal fluctuations while satisfying balancing constraints: regional quarterly sums must match the annual regional data, and the sum of regional quarterly values must equal to the national quarterly figure. The resulting optimization problem is solved via an interior-point algorithm. Its sequential variant requires revising only the most recent disaggregated series when data for a new year become available. A regression model with four benchmarks is used for comparison.

Results

The sequential smoothing method produced quarterly regional GDP growth rates with a level of smoothness comparable to national data while still capturing local variations (e.g. the COVID-19 downturn). A comparison with the one-step version confirmed almost identical results. Against the regression-based approach, average differences in annual growth rates reached up to 0.8 percentage points, with correlation coefficients exceeding 0.87. Forecasting performance analysis indicated that the smoothing method is competitive with regression when series are short or horizons are brief.

Conclusions

The proposed procedure is computationally simple, requires no auxiliary indicators, and can be implemented by statistical offices lacking long, high-quality regional data series. Its sequential variant minimizes historical revisions while maintaining close agreement with regression-based methods. Combining smoothing with a regression-based approach could be considered for further improvement.

Keywords

regional GDP, temporal disaggregation, nowcasting, smoothing algorithm, interior point optimization

Objective

For many years, the public debate in Poland has highlighted the issues of urban sprawl and spatial chaos. While various monitoring datasets exist, there is a lack of nationwide, high-resolution spatial statistics. Available land use data often fail to capture the functional aspects of urban areas. The aim of this study is to identify and analyze the functional land use structure of Polish cities using a consistent methodology based on official spatial data, as well as to examine changes between 2017 and 2024.

Methods

The research method is based on the concept of the urban transect. Functional land use analysis was conducted using data from the Land and Building Register (EGiB), with 500 m2 hexagons as the basic unit of analysis. Based on land use structure, the hexagons were classified according to a defined functional typology. To reflect the specific characteristics of Polish cities, two additional categories were introduced: residential development on agricultural land and high-density multifamily housing areas.

Results

The conducted analyses allowed for the identification of the functional and spatial structure of cities, enabling an objective characterization of urban land use and an assessment of its temporal changes. Based on the results, it is possible to broadly identify spatial reserves within cities for potential future development and to indicate the likely directions of urban expansion. The findings also revealed the intensity of uncontrolled and chaotic suburbanization processes.

Conclusions

Due to its high level of automation, ongoing data updates, and relatively easy access, the proposed research method offers a viable alternative to traditional studies of the functional and spatial structure of cities. The resulting data – accurate yet standardized and generalized – effectively reflect changes occurring in urban space, making them a valuable source for analyzing the evolution of spatial structure over any selected time period.

Keywords

urban space, LULC, EGiB, GIS

Objective

This study proposes a new density-based metric for measuring spatial competition among urban retailers. It integrates population distribution and firm locations using Voronoi tessellation to define zones of influence. Inspired by ecological competition models and spatial econometrics, the method addresses limitations of traditional approaches and supports practical applications in urban planning, retail strategy, and antitrust analysis.

Methods

Retail locations from OpenStreetMap are cleaned using spatial clustering and medoid selection. Voronoi polygons define each facility`s influence zone within Warsaw. Population data from the Polish Census is proportionally assigned to each zone. Population density serves as a spatial weight to compute standardized market share indices and a spatially weighted Herfindahl-Hirschman Index (HHI).

Results

Applied to Biedronka and Lidl supermarkets in Warsaw, the method reveals that densely populated areas yield higher market influence per store. Spatially weighted HHI values differ notably from traditional metrics, exposing deeper patterns of market dominance. Results show the model’s ability to correct for spatial distortions in raw data and better reflect real competitive dynamics.

Conclusions

The proposed metric offers a replicable, data-driven tool for evaluating competition in urban environments. It improves the understanding of firm influence, supports equitable access assessments, and enhances decision-making for retail and urban policy. The method is adaptable to other sectors and cities using open data, and bridges ecological theory with spatial economic analysis.

Keywords

spatial competition, Voronoi tessellation, urban retail, market concentration, population-weighted index

Objective

The aim of the paper is to present the assumptions of the Integrated Analytical Platform - a planned platform to support the initiation and implementation of research using data from public records.

Methods

The paper presents the results of the 'Integrated Analytical Platform' project, implemented in partnership between public institutions (Minister of Digitalisation, Minister of National Education, Minister of Family, Labour and Social Policy, Minister of Health) and academic institutions (University of Warsaw, SGH Warsaw School of Economics).

Results

The result of the project is a prototype of an analytical solution supporting conducting analyses on the basis of combined data from different public registers: the main assumptions of the methodological approach to the use of ZPA as a tool for cooperation between public administration and research centres are presented.

Conclusions

Conducting analyses based on data from public registers is key to building evidence-based knowledge. Pilot research within the Integrated Analytical Program project demonstrates the potential of using administrative data to conduct analyses of public policy instruments.a

Keywords

administrative data, evidence-based policy

Objective

The aim of the presentation is to present an original QDC (Quick Density Clustering) algorithm, which is used to divide point data into groups according to spatial density. This is a new solution that complements the existing tools for spatial quantitative modelling, based on spatial machine learning methods.

Methods

The algorithm is based on a highly intuitive design that exploits the characteristics of varied spatial density. Points located close to each other can be easily extracted because the sum of distances to the k nearest neighbours is low, while the number of neighbours within a certain radius is high. Conversely, points from low-density areas have a high sum of distances to the k nearest neighbours, while the number of neighbours within a certain radius is low. These two features clearly characterise areas with low and high density.

Results

Dividing these two spatial variables (normalised) using the k-means algorithm into k groups allows geo-localised points to be divided into groups with similar spatial density in a simple and quick manner. The application of the QDC algorithm to data on human activity – the location of residents (data from the census) and the location of companies (data from REGON) – will be presented. Relative spatial densities will be compared based on data rasters and in relation to the road and settlement networks downloaded from Open Street Map.

Conclusions

The QDC algorithm is an interesting methodological innovation that has many applications in monitoring social mobility dynamics. It is a sampling-resistant solution that is also easily adaptable to streaming (the constantly growing) data.

Keywords

spatial density clustering, geolocated point data, spatial machine learning

Objective

This study investigates the core-periphery pattern in public transport accessibility within the Warsaw Metropolitan Area. It aims to quantify and analyze the spatial disparities in accessibility, as well as time-based service fluctuations. The study allows the evaluation of the efficiency of transport services in facilitating access to economic opportunities, particularly employment.

Methods

A gravity-based model was applied to quantify accessibility, using two complementary measures. The first utilized population size within a 1×1 km grid as a proxy for destination attractiveness, while the second assessed accessibility to jobs at the municipal level, adjusting for competition and transport mode share. Travel times were calculated using the r5r routing engine with input from GTFS schedules and OpenStreetMap road networks. Time-decay functions were empirically derived from travel survey data, and both spatial and temporal variability were analyzed.

Results

The findings reveal a highly centralized accessibility pattern, with the highest accessibility in Warsaw`s city center and sharp declines in outlying areas. Accessibility is strongly correlated with proximity to railway lines and inversely with travel time to the center. Temporal analysis highlights fluctuations in service levels, with suburban areas experiencing greater variability and lower accessibility during off-peak hours and weekends. A standardized accessibility gain index, measuring the contribution of public transport over walking, confirmed the core-periphery disparity.

Conclusions

The study confirms a core-periphery structure in public transport accessibility across the Warsaw Metropolitan Area, with significant temporal and spatial inequalities. These disparities suggest a need for targeted investments in peripheral areas to ensure equitable access to employment and services. Methodologically, the thesis introduces a novel normalized accessibility gain index and validates the potential use of simpler cumulative opportunity models. Future research should consider broader definitions of destination attractiveness.

Keywords

public transport accessibility, gravity model, Warsaw Metropolitan Area, urban planning

Objective

This study addresses the question of whether technological relationships are asymmetric, and how certain technologies act as drivers of innovation within regional clusters. The main objective is to develop a methodological framework that captures directional (asymmetric) relationships between technologies, with the aim of identifying those that function as catalysts in regional innovation systems.

Methods

The research uses an unsupervised machine learning approach combining two techniques: DBSCAN (density-based spatial clustering) to group patents into spatial clusters, association rule mining (a priori algorithm) to uncover asymmetric co-occurrence patterns between technology types over time. The dataset includes UK patent applications filed between 1980 and 2015, sourced from UKIPO and PATSTAT. Each patent is geocoded using inventor coordinates and classified by technology type (high-tech vs. low-tech), and grouped into 5-year time intervals.

Results

The study showed that: technologies are asymmetrically related, high-tech patents are influencing low-tech more, compared to the opposite, and H04Q (selecting), H04N (pictorial communication e.g. television), H04M (telephonic communication) and H04L (transmission of digital information, e.g. telegraphic communication) affect each other very frequently,forming a 'technological cluster'.

Conclusions

This study contributes to regional science and innovation studies by providing a novel methodological framework for detecting asymmetric technological relationships. It challenges the common assumption of symmetry in relatedness measures and offers practical tools for identifying catalyst technologies that may support regional diversification strategies. The method is generalizable and can be applied to any spatio-temporal dataset with classification variables, making it valuable for broader applications in policy analysis, technology forecasting and economic geography.

Keywords

asymmetric technological relatedness, patent clustering, regional innovation, catalyst technologies, spatio-temporal analysis

Objective

In our paper we aim to analyse the geographical patterns of the spread of Polish voters’ tolerance for parties’ populist rhetoric and anti-EU position between the lower chamber elections in 2019 and 2023. Our particular focus is on the effect of the outburst of the war in Ukraine in 2022 on the spread of the abovementioned attitudes. We aim to provide the answer to the following research question: Is the municipality proximity to the Polish-Ukrainian border correlated with voters’ tolerance for populist and anti-EU attitudes of political parties? What is the sign of a correlation?

Methods

We apply here standard spatial regressions from the simplest ones (as spatial error model) to the most general (Manski model), as mentioned by Elhorst (2010). Moran’s I was calculated before running spatial regressions. Pre-estimation started with Anselin’s (1988) algorithm, in order to determine which models have to be estimated. Post-estimation is presented by the Hausman test (Pace & LeSage, 2008), ANOVA and information criteria. Our data come from various source, like PKW, GUS, GUGIK, the 2019 and 2023 Chapel Hill Expert Survey

Results

Location of the municipality in a border zone is significantly correlated with the change in the intensity of voters’ tolerance for parties’ anti-EU position. The voters from these municipalities had lower tolerance for parties expressing anti-EU positions in 2023 as compared to the previous elections. We hypothesise that this effect is due to the higher threat stemming from the proximity to war-torn territory. However, once we account for the alternative distance measure calculated from the centroid of the municipality to the closest point of the border, we find a significant negative relation.

Conclusions

Our results highlight the importance of the exposure to the war in Ukraine as one of the factors correlated with the increase in Eurosceptic views and tolerance towards populism in Polish municipalities. We believe that they may be valuable for country- and EU-level policymakers. What is more, we see a room for future research on the impact of proximity to the war in Ukraine on a variety of other social, political, and economic phenomena.

Keywords

voters behaviour, spatial modeling, spatial autocorrelation, distance measures, populism, Euroscepticism

Objective

The aim of the paper is to propose a procedure on how to use multidimentional scaling in a way that makes it possible to analyse the dynamics of comparable results

Methods

There are n objects characterized by m variables. MDS places these objects in lower dimensional space (m*), usually on a plane – in a way that distance matrix calculated in m* is the closest to the Dm. Axes in the final configuration have no interpretation and the mean position of objects is (0:0). We propose to rotate the configuration in year t in such a way that the sum of the squares of the distances to the configuration for t-1 is minimal. Thus, the step-by-step rotations give comparable results which are subject for trend estimation, analysis and prediction.

Results

An illustrative example deals with 27 European Union countries, characterized by GDP, expenditures on R&D, employment, inflation and innovation, within the 2014-2023 period. Trends and forecasts are presented.

Conclusions

The coordinates of a system where MDS results are presented, has axes with no interpretation, so the results of the same objects characterized by the same set of variables can differ in some random way from year to year. The proposed rotation leads to comparable results which can be smoothed by analytic trends, and give forecasts of MDS configuration.

Keywords

multidimensional scaling, rotation, trends, EU countries

Objective

This paper addresses the conceptual and practical challenges of applying artificial intelligence (AI) and spatial machine learning (SML) in regional science. The goal is to clarify how SML differs from more general applications of AI in the spatial context, and to demonstrate their value for solving real-world location-based problems. We argue that AI offers a novel analytical framework that complements traditional methods and leads to more accurate, data-driven policy and business recommendations.

Methods

Our approach combines spatial machine learning (SML) with an interactive AI-based application designed to support decision-making. We use data on business entities in Poland, enriched with detailed spatial characteristics of their environment. SML algorithms are employed to detect spatial patterns and predict optimal business locations. These results are then embedded into the AI app, which enables users, such as policymakers and entrepreneurs, to explore scenarios and receive recommendations tailored to specific regional contexts.

Results

Our results show that spatial machine learning (SML) offers strong predictive performance in identifying optimal business locations. However, AI extends beyond prediction, enabling the generation of strategic recommendations for both business and policy. This distinction is crucial: while SML identifies patterns in spatial data, AI provides a broader decision-making framework by integrating predictions with contextual reasoning and actionable outputs. This combination supports more informed regional planning and business choices.

Conclusions

This work contributes to the emerging literature on AI in regional science by offering clear definitions and practical examples that distinguish spatial ML from broader AI applications. The proposed model serves both researchers and practitioners by demonstrating how AI can support location-based decision-making. Our findings highlight the need for a better integration of spatial thinking into AI workflows, ultimately improving regional policy design and business strategies.

Keywords

Artificial Intelligence, spatial machine learning, regional science, business location, spatial data

Objective

In this article, the authors explore the complexity of integrating diverse datasets with different distributions. They emphasize the need to identify the underlying distribution of the data in order to select the most appropriate analytical methods.

Methods

The proposed methodology starts with evaluating data sets for a power-law distribution using statistical tests such as the Gabaix or Kolmogorov-Smirnov tests. If the data support this distribution, the authors recommend using the generalized cross-entropy technique associated with Tsallis, which is particularly useful in capturing the complexity of non-ergodic systems plausibly associated with Big Data. On the other hand, if the data sets do not exhibit the characteristics of a power-law distribution, the paper proposes traditional econometric methods, including the Shannon cross-entropy.

Results

By providing a structured framework for analysis, the authors contribute valuable insights that increase the quality and credibility of the results obtained from Big Data.

Conclusions

The work outputs, illustrated by a few case studies, should help avoid the intemperate misuse of models based on normality assumptions, currently often expanded to the case of Big Data.

Keywords

data integration, Big Data, complex systems, power law, cross entropy econometrics

Objective

In this presentation, we will recall the basic assumptions of the Statistical Disclosure Control algorithm prepared for the needs of the Geostatistical Portal of Statistics Poland, one of the basic and most comprehensive channels for sharing and analyzing Polish official statistics data. We will present the assumptions and the scheme of the optimal data protection process and indicate the basic problems related to the practical implementation and the possibilities of solving them.

Methods

In the research, we use an algorithm written in SAS and R using the sdcMicro package of the R environment, its application and parameterization.

Results

We will discuss the effects of applying the algorithm and correcting the obtained disturbances in order to maintain the definitional consistency of the provided data and the summation of appropriate values.

Conclusions

We will point out the usefulness of the algorithm and outline the most important challenges facing official statistics in this area – both methodological (differential privacy, generating synthetic data) and organizational (e.g. access to data for scientific purposes).

Keywords

Statistical Disclosure Control, Geostatistical Portal, perturbative methods, synthetic data

Objective

Forecasting exchange rates is a key issue in finance and economics. The purpose of the conducted study was to evaluate the effectiveness of graph neural networks (GNNs) in forecasting and classifying returns for exchange rates, with a special focus on the structural relationship between currencies.

Methods

The analysis included time series of returns for 15 major world currencies over the 2022-2024 period. The structure of the network was determined based on correlations between returns, using minimum spanning trees (MST), in which nodes represented individual currencies and edges reflected the interdependencies between them. Forecasting and classification of returns for exchange rates was performed using convolutional graph networks (GCNs). The results were compared with benchmark models using ex-post forecast errors (RMSE, MAE) and measures of classification accuracy (Acc).

Results

Traditional models, such as ARMA, for example, analyze time series independently, ignoring the interconnectedness between currencies. However, exchange rates form a dynamic network of relationships that can be effectively modeled using GNNs. The constructed graph networks took into account the structural relationships between currencies. Both the GCN and LSTM-GCN models outperformed traditional statistical methods in terms of prediction and classification quality. The analysis of embeddings of nodes made it possible to identify the most influential currencies shaping the market structure.

Conclusions

The growing interest in GNNs in financial data analysis in recent years suggests their potential superiority over classical statistical methods. The results obtained confirmed this, and the built predictive graph neural network models can support investors and financial institutions in decision-making and risk management.

Keywords

graph neural networks, exchange rates

Objective

The purpose of the empirical study presented in the paper is to compare linear discriminant models constructed on the basis of diagnostic variables selected using different techniques for their selection, and to indicate which bankruptcy prediction model proved to be the best from the point of view of the multivariate evaluation of model accuracy.

Methods

The study applied a quantitative approach using discriminant analysis to identify companies at risk of bankruptcy. The effectiveness of the models was evaluated using Wilks` lambda statistic, AUC values and classification accuracy of the models. A number of methods were used to select discriminating variables: arbitrary selection of variables based on the literature and using GPT Chat (version 3.5), one-way analysis of variance (ANOVA) and stepwise forward selection, Hellwig`s method of selecting diagnostic variables, t-statistics, backward stepwise method. The source of the data was the EMIS

Results

Eight discriminant models based on different methods of selecting diagnostic variables were evaluated. The highest classification efficiency (both in the learning and test sample) was achieved by model D, built using a two-step method (significance test of mean differences + progressive selection). Models F (based on isolated variables selected by the Hellwig method) and H (backward method) also had high classification efficiency. In contrast, the worst results were obtained by model A, based on arbitrary selection of variables. AUC indices for most models (except for model A) exceeded the val

Conclusions

The results of the study indicate that appropriate independent variable selection techniques play an important role in the process of building discriminant models. The study also makes an important methodological contribution, challenging the common belief that Wilks` lambda statistic is highly useful in assessing the quality of classification models. The results show that the analysis of classification performance should be the main point of reference when evaluating discriminatory models.

Keywords

diagnostic variable selection, linear discriminant function, bankruptcy prediction, Euclidean distance

Objective

The purpose of the study is to define a representative based on the available data. Validation of the transaction data set from selected travel agencies and indication of the characteristics of the representative are the stages of the analysis. They are the starting point for calculating the price index for tourism.

Methods

The source data comes from the administrators. The challenge in this study is the verification of the transactional data set. It results from the non-uniform characteristics of information about products / offers obtained from data providers. The characteristics of individual offers differ in the number of categories and specifications. The high variability of the product offer over time results from the nature of the data, which is characterized by seasonality over a period of one year. The problem is also the lack of data in the time series – seasonality for a given offer.

Results

The statistical analysis carried out, both with the use of statistical tests and an attempt to build and estimate the hedonic model, made it possible to indicate potential representatives for the calculation of the price index.

Conclusions

When setting the threshold of importance for the weights from the data providers, it is possible to indicate a representative associated with a group of tourist offers for the discussed direction or destination. Preliminary calculations of the price index of tourist services indicate a high volatility of its value.

Keywords

price index, weight for offer, weight for representative, destination, hedonic

Objective

The aim of the study is to provide a description of the actions that statisticians should take in order to, on the one hand, more fully utilize the potential of statistical inference in the era of big data and AI and, on the other hand, to make statistics users more aware of the inherent elements of uncertainty present in statistical inference.

Methods

The method used is a critical analysis of the literature and research studies that contain elements of statistical inference. A creative reflection on the practice of statistical research in the context of the groundbreaking work of Ronald Fisher Statistical Methods for Research Workers” published in 1925 was also mentioned.

Results

Discussions on the dilemmas related to the category of statistical significance held in the scientific community worldwide is a consequence of the incomplete consideration of model assumptions in the practice of statistical inference and of assigning a decisive significance to a single sample study in the context of the universal goals of scientific research.

Conclusions

With respect to the practice of testing statistical hypotheses, we postulate the use of the the Bayesian approach as complementary to the classical approach inextricably linked to Ronald Fisher.

Keywords

statistical inference, uncertainty, model assumptions, big data, scientific research

Objective

The aim of the study is to identify and evaluate the influence of key factors (TIVHE) on the quality of customer retention classification models and to compare the effectiveness of econometric and advanced machine learning methods.

Methods

The study is simulation-based aligned with the CRISP-DM methodology using a complex customer retention dataset. The influence of five factor groups was analyzed across logistic regression, decision trees, neural networks and ensemble models.

Results

The findings of the study show what follows: predictor transformation was most beneficial for logistic regression, feature engineering and hyperparameter tuning improved all models, segmentation helped interpretable models but harmed black-box methods, and aggregation improved predictions for all except logistic regression.

Conclusions

Effective model development requires aligning the TIVHE strategy with the nature of the algorithm used. The results underscore the need for a flexible modeling approach, offering guidance for enhancing model performance.

Keywords

machine learning, data mining, hyperparameter optimization, classification models, customer retention models, model ensembles

Objective

The subject of the study is to present the methodology for estimating changes in taxi service prices in Poland using data collected from the websites of companies and taxi corporations (i.e. using the so-called web-scraping). One of the objectives is also to present the technological and methodological challenges that the statistical office faces when it wants to implement this type of data in the regular production of price indicators. The presented solutions are, among others, the result of activities carried out as part of a project financed by the European Commission.

Methods

The subject of presentation is to present the methodology for estimating changes in taxi service prices in Poland using data collected from the websites of companies and taxi corporations (i.e. using so-called web-scraping). One of the objectives is also to present the technological and methodological challenges that the statistical office faces when it wants to implement this type of data for regular production of price indicators. The presented solutions are, among others, the result of activities carried out as part of a project financed by the European Commission.

Results

As a result of the work carried out, a method for calculating price indices for taxi services in Poland was developed, as well as the tools necessary to collect and gather data from carriers' websites.

Conclusions

The main conclusion from the undertaken action is to implement the methodology for estimating the dynamics of taxi service prices developed under the presented work into the statistical production. It is also recommended to use an automated method for collecting data made available in the public space, which will contribute to an increase in the number of observations of changes in the levels of taxi service prices, as well as to expand the range of service representatives. In addition, solutions and experiences gained in the area of price statistics should be used in other areas of statistical surveys.

Keywords

taxi service prices, stripped data, elementary index, Dutot index

Objective

The aim of the paper is to present and analyze the development of scientific research and teaching work conducted by statisticians and demographers from Łódź.

Methods

The presentation of the achievements of statisticians and demographers from Łódź is based on an in-depth analysis of the literature on the subject and the study of documents and source materials.

Results

Based on the analyses, it can be stated that the statistical research conducted in Łódz in the interwar period had a significant impact on the development of Polish statistics. In particular, these were works conducted under the supervision of the head of the Department of Statistics of the Łódź City Hall, Edward Rosset, and professor Edward Szturm de Sztrem, which addressed important socio-economic problems of Poland and Łódź, including: 'Alcoholism in the light of statistical researc': 'The political face of the population of the city of Łódź in the light of election research'.

Conclusions

The interwar achievements of Łódź statisticians continued after Poland regained independence within the framework of the University of Łódź, the Łódź branch of the Warsaw School of Economics and the Higher School of Economics. The achievements of the Łódź statistical community, which took place over the last 100 years, contributed to the creation of the Łódź school of statistics and demography.

Keywords

statistical research, development of statistical thought

Objective

The aim of the paper is to present, on the basis of all currently available archival source data and historical studies, the situation of the Polish industry in the General Government in the years 1939-1945.

Methods

The study uses historical analysis based on the available printed and archival sources.

Results

the work has a pioneering character

Conclusions

the first time in the literature the topic was taken up

Keywords

history of statistics, World War II, General Government, industry

Objective

The study examines the origins of demographic forecasting in post-war Poland.

Methods

The study is based on an analysis of studies published in the 1950s devoted to attempts at demographic forecasting.

Results

In the 1950s, attempts to create demographic forecasts for Poland in general were undertaken in three public centres: the Institute of Urban Planning and Architecture operating at the Ministry of Urban and Housing Construction, the Central Statistical Office and the Planning Commission at the Council of Ministers. Of these centres, the last two cooperated closely with each other.

Conclusions

Comparing the seven projections prepared in the years 1953-1961, one can notice: 1) a departure from single-variant projections to multi-variant ones, coexisting with a departure from the assumption of constant birth, death and internal migration rates, 2) transformation of projections prepared for Poland with a division into the population of cities and villages into projections referring to smaller administrative units – voivodships and cities with voivodeship rights.

Keywords

demographic forecasts, Poland, 1950s demographic forecasts

Objective

The aim of the paper is an attempt to outline the evolution of the income distribution of Poles in the period from the Second Polish Republic to the present day. Particular attention is paid to changes in the concentration of income, the synthetic indicator of income inequality, the indication of the income gap and not the level or value of income.

Methods

Sources of income information for the period up to the end of World War II, mostly non-statistical, are scattered and fragmentary. A method based on the analogy of income distribution in countries with a social and economic situation similar to Polish will be used. Data for the period of the People`s Republic of Poland are incomplete. Using the stylized facts approach, the likely magnitude of Gini`s income concentration indicator will be estimated. The periods of analysis include the interwar period, the German occupation, the People`s Republic of Poland and the Third Polish Republic.

Results

In each of the separate periods of the analysis, income inequality evolved for different reasons and the spatial differentiation of income is still influenced to some extent by the effects of partitions. The log-normal distribution, i.e. the natural shape of the income distribution, accompanied the market economy in the interwar period and after 1989. Between 1939 and 1989, the natural distribution of income was distorted.

Conclusions

As expected, the distribution of Poles` incomes over the past 100 years has largely depended on the economic system and significant historical upheavals, such as World War II, the change of state borders and population movements, the continuation and overthrow of the centrally planning system.

Keywords

income distribution, concentration of income. income inequality, income gap

Objective

Statistics is often perceived by students as a difficult, abstract and rather unpleasant subject. The aim of the activities I will discuss in this presentation was to increase student engagement and improve the understanding of key statistical concepts through the use of active learning methods.

Methods

As part of a statistics course for students of economic analytics, a range of activities inspired by English-language educational literature was implemented (including Gelman & Nolan, 2017: Schaeffer et al., 1996). Everyday objects (a tape measure, kitchen scale, beach ball, candies, and dice), simulation games, and playful elements (such as creating memes) were used to illustrate both basic and advanced tools of probabilistic modeling and methods of statistical inference.

Results

The introduction of these activities increased student interest in the subject and made it easier to grasp difficult topics. Although the final exam remained challenging, 90% of students passed it on the first attempt. The course was also rated more highly than in previous editions.

Conclusions

Using active learning methods based on simple props and interactive forms of work can significantly improve the effectiveness of teaching statistics. This approach allows students to gain a deeper understanding of the material and enables the instructor to observe genuine engagement and those “aha!” moments during class. Effective and well-organized collaboration among academic instructors, including the sharing of proven teaching strategies, contributes to improved teaching quality, increased student engagement, and the broader popularization of statistics.

Keywords

statistics education, active learning, teaching innovations

Objective

The aim of this paper is to present the original educational method 'Know thyself', which employs statistics to develop students’ research competencies and self-awareness. The research question investigates economics students` awareness of the fake data problem and examines how effectively they detect and interpret manipulated data within practical and ethical contexts.

Methods

A didactic experiment replicating canonical research on randomness perception was conducted. Students generated fictional accounting datasets subsequently analyzed using Benford’s law. Additionally, a survey assessed students’ awareness of fake data, their previous experiences, and strategies for handling data manipulation.

Results

Findings revealed students` limited proficiency in identifying fake data, confirming the importance of developing analytical and critical competencies. Simultaneously, increased student engagement was observed when exercises highlighted real-world consequences of misinformation and ethical dilemmas associated with data reliability.

Conclusions

The results indicate the necessity of integrating data verification and ethical responsibility topics into curricula. The 'Know thyself' method enriches statistical education by incorporating philosophical and ethical aspects, enabling students to critically evaluate data and their own decisions, thus contributing significantly to the didactic literature and educational practice in economics.

Keywords

Know thyself method, fake data, Benford’s law, teaching microeconomics, ethics in statistics

Objective

Statistical education is of great importance in the modern world: it influences both the worldview and the decisions a young person makes. School is the first environment where students learn about statistics and its importance theoretically and practically, and not only during mathematics lessons. The aim of the study is to present examination tasks from the last 10 years, where statistical knowledge and skills proved crucial to their solution.

Methods

The presentation is based on an analysis of examination papers, analysis of sources and publications, as well as interviews.

Results

The analysis of examination papers in science and social science subjects shows the universality and functionality of statistics.

Conclusions

The importance of disseminating statistical knowledge and skills is not always emphasized in the school setting. It is worth emphasizing, however, that the Statistical Olympiad organized by Statistics Poland and the Polish Statistical Association, which will have its 10th edition in 2025/2026, is enjoying increasing interest and emphasizes the importance and significance of statistics for young people.

Keywords

mathematical statistics, statistical education, universality and functionality of statistics, Statistical Olympiad

Objective

The purpose of the study was to determine perceptions, experiences and attitudes towards sustainability micro-statements in academia, and to explore their potential as a tool to support education.

Methods

The study was exploratory in nature. Its goal was to understand the perception and potential of micro-statements as a tool to support sustainability education in academia. The online questionnaire targeted students at public universities in Poland.

Results

Respondents showed moderate familiarity with microcredentials, but at the same time high interest in their teaching potential, especially in the context of flexible learning and rapid response to labor market needs. The majority of respondents considered microcredentials to be an important or very important tool to support the development of competencies in social justice, green finance and the circular economy. The data collected also confirms the importance of interdisciplinarity, active teaching methods and cross-sector cooperation (university-business-administration).

Conclusions

Micro-credentials are seen as a valuable tool for education and communication, especially in the field of sustainability. They increase student engagement and help align competencies with labor market requirements. However, the implementation of microcredentials requires better staff preparation and integration with university programs. Universities should support their development, taking into account various forms of teaching and cooperation with the external environment.

Keywords

microcredentials, sustainable development, higher education, students

Objective

The presentation aims to show how modern AI-based technologies can support organizations in internal and external communications, with a particular focus on activities in the areas of accessibility, HR and employee psychological support. The paper will discuss the existing technological solutions, their practical applications, as well as potential ethical and organizational challenges associated with their implementation.

Methods

The presentation will present the practical application of AI in the areas identified in the research objective. Possible technologies, benefits and risks will be discussed, as well as stages and an implementation scenario. The aspect of ethics in AI implementation will be addressed. Examples of the use of appropriate prompts to facilitate the use of AI resources will be mentioned.

Results

In the era of digital transformation, institutions and business units face the challenge of adapting to the requirements and expectations of external as well as internal audiences. Also, statistical offices, in order to meet the requirements and remain competitive, should and even are obliged to keep up with the intensity of technological change, while keeping in mind the sensitivity of audiences, especially from groups with special needs.

Conclusions

Mimo wielu korzyści płynących z korzystania z dostępnych narzędzi AI, należy mieć jednocześnie na uwadze fakt, iż technologie te wiążą się z ryzykiem naruszenia prywatności, nieprzejrzystością algorytmów oraz ograniczonym zaufaniem użytkowników. Konieczne jest więc etyczne i odpowiedzialne podejście do ich projektowania i wdrażania.

Keywords

AI, employee mental health support, HR, accessibility

Objective

Artificial Intelligence enters into official statistics. However, it is not a miraculous invention that makes every work immediately perfect. To cooperate with AI, you need to have expert knowledge in the given field of statistics, have a vision, the ability to write prompts, algorytms and know how to combine them synergistically to obtain measurable effects and avoid the traps of AI hallucinations. How to do this is the subject of the paper.

Methods

The methodology of the approach to work with AI was based on the principles, methods and techniques presented in the publication 'Modern technologies and new data sources in inflation measurement' from 2022, edited by Jacek Białek, Mieczyslaw Klopotek and Tomasz Panek. The paper presents practical issues related to the use of AI techniques in inflation measurement, with attention paid to the occurrence of certain barriers and limitations affecting the quality of statistical processes and obtained statistical data.

Results

The paper discusses the challenges that exist in the area of trade and services. It defines what AI is and how it can help in the conducting of statistical surveys. It presents the advantages and disadvantages of AI application when using scanned data and scrapped data in measuring CPI/HICP inflation. Among the factors influencing the quality of work with AI, attention was paid to, among others, the role of critical thinking and its importance in achieving statistical goals.

Conclusions

Thanks to AI, we can use the data potential in a more optimal way from a diverse sources ecosystem. However, the key issue is always the relevance of these data to the construction of strategies or state policies.Therefore, each result that we obtain from the work with AI must be assessed by a statistician in terms of its consistency with methodological and qualitative requirements of the standards of official statistics.Such a rational approach can protect us from production of artificial results by AI that we are unable to interpret on the basis of available statistical knowledge.

Keywords

Artificial Intelligence, machine learning, data scanning, data scraping, critical thinking

Objective

The aim of the paper is to draw attention to the non-existence of a separate code of professional ethics in force in Polish official statistics. Relying on international practices, i.e. standards and aspects of statistical ethics contained in the European Union`s Code of Statistical Practices and the Fundamental Principles of Official Statistics of the United Nations, seem to be insufficient. They are also not properly separated.

Methods

The methods used are based on the analysis of the content of the Code of Statistical Practices of the European Union and the Fundamental Principles of Official Statistics of the United Nations in terms of distinguishing the elements of these documents that relate to the issue of ethical principles in official statistics. Additionally, a critical analysis is conducted, aimed at pointing out certain deficits in international standards of ethics in statistics.

Results

The analysis will include a review of the practical application of basic ethical principles in official statistics. Cases of conscious or unconscious violation or circumvention of ethical principles will be cited. The risks associated with the possibility of violating these rules will be discussed. The consequences of the lack of a sense of public importance of statistical information by its creators – as a public good – for its recipients will be indicated.

Conclusions

The conclusion of the paper is the postulate to start a discussion and then work on the creation of a code of professional ethics in official statistics in Poland. Following the example of the practice and experience of several countries, I believe that the right place to develop the postulated code may be a team of independent experts or the Polish Statistical Association.

Keywords

ethics in official statistics, codes of professional ethics

Objective

In an era of information overload, statistical data must compete for attention with emotions, narratives and... memes. So how do we effectively talk about the labour market in a language that not only conveys the complexity of the phenomena, but also reaches the audience?

Methods

This presentation is a journey through the world of numbers, charts and statistical interpretation – in search of ways to make statistics not ‘dry’ but attention-grabbing and understandable. I will talk about the challenges of data communication: from misinterpretations to oversimplifications. But also about opportunities – new tools, visualisations and data storytelling techniques.

Results

Effective communication of statistics is not just a matter of form – it is the foundation for better decisions, policies and actions in the labour market.

Conclusions

For the presentation of activities, I will use materials and activities prepared by the Statistical Office in Bydgoszcz to promote the results of the Survey on Employment in the National Economy and the Survey on the Distribution of Wages and Salaries in the National Economy, but also various forms of media communication of the data made available.

Keywords

labour market, statistical data, data communication, data visualisation, data interpretation, data user, communication challenges, data presentation tools

Objective

The aim of the paper is to present the characteristics of relative indicators for monitoring the labour market entry of graduates used in the Polish ELA system, from the perspective of their application for comparing the entry of graduates into the labour market over time (individual cohorts) and in space (locally, but also internationally).

Methods

The analyses use data from the ELA system, based on a combination of information from two public registers: POL-on and ZUS (Social Insurance Institution), as well as data from graduate monitoring systems in selected countries (the Czech Republic, Slovakia, Hungary, Romania). Relative earnings ratios compiled from data from these systems are comparable.

Results

The obtained data confirm the good properties of relative employment rates for comparisons over time and space. Moreover, these indicators are internationally comparable, as confirmed by pilot analyses conducted using graduate tracking systems in 4 European countries.

Conclusions

The innovative relative indicators used for the first time to monitor the higher education graduates in Poland have characteristics that allow them to be used for international, local and also temporal comparisons.

Keywords

graduates tracking, monitoring employment, relative earnings

Objective

A comprehensive framework identifying and examining the multifaceted problems (technical, governance-related, legal, ethical, economic, operational, socio-cultural) impeding the successful development, deployment and sustained operation of public sector data spaces, and to understand their interdependencies and collective impact is the aim of the presentation.

Methods

The study is based on an extensive review and synthesis of the academic literature, policy documents, government reports and case studies of European public sector data space initiatives. It utilizes a qualitative approach, analyzing documented challenges and lessons learned to build a comprehensive taxonomy and analytical framework of problems.

Results

The key findings reveal a wide spectrum of interconnected problems in building public sector data spaces. These span technical, governance, economic and socio-cultural dimensions, often creating compounding negative impacts.

Conclusions

The research provides a comprehensive taxonomy of the challenges, informing policy and practice for more effective public sector data space implementation. It underscores the need for holistic strategies addressing technical, governance, economic and socio-cultural issues to realize public value, foster innovation and retain citizen trust. A future research agenda is proposed to tackle unresolved problems and guide sustainable development.

Keywords

data spaces, public sector, data governance, European data strategy, digital transformation, data sharing, legal frameworks

Objective

In the face of dynamic social, economic and technological changes, official statistics require not only tools to respond to the emerging information needs but, above all, the ability to anticipate those needs and turn them into new development opportunities. This necessitates the development of a culture of innovation. The aim of this presentation is to outline the ways in which the innovation culture is being fostered within official statistics, and to indicate directions for further action.

Methods

Using a systemic approach, we create internal conditions for testing new solutions, support teams experimenting with alternative data sources and novel analytical methods (including AI, big data and the integration of administrative data), and promote cross-sectoral cooperation – with the private sector, academia and public administration – as well as international collaboration. We understand innovation not only as the adoption of new technologies but, above all, as an attitude of openness to change, learning and continuous improvement.

Results

Statistics Poland's activities contribute to strengthening the capacity of official statistics to test new solutions, make better use of data and enhance collaboration with stakeholders. There is a noticeable increase in awareness and organizational readiness to implement change and openness to new approaches. It is essential that innovative thinking is embedded not only in analytical and methodological efforts but also in everyday interactions with data users.

Conclusions

Building a culture of innovation in statistics is a long-term process that requires a systemic approach, support for experimentation, and both internal and external collaboration. Statistics Poland's experience shows that innovation can be a lasting source of value in statistics – methodologically, organizationally, and socially.

Keywords

innovation, strategy, cooperation

Objective

The aim is to present the possibilities of calculating price indices and rents for commercial real estate (retail, office, warehouse) in Poland.

Methods

The methodology usually takes into account the various types of indices calculated, mainly hedonic: time dummy, rolling time dummy, repricing, average characteristics, imputation, stratification method.

Results

The choice of the method for estimating the index depends on the amount of statistical information available in the database. More extensive databases, which include information such as property characteristics, allow for the calculation of hedonic indices, for example, using the rolling time dummy method.

Conclusions

Improving the quality of rental and commercial property price indices requires obtaining additional sources of data on commercial properties (other than the Property Price Register).

Keywords

price indices, rent indices, commercial real estate, hedonic indices

Objective

Primary health care (PHC) is the cornerstone of the health care system and plays a key role in ensuring the availability of medical services, the coordination of patient care and the implementation of preventive measures. The aim of the presentation is to illustrate the importance of the PHC by analysing selected statistical data on e.g. consultations in the PHC, quality of services as assessed by patients, accessibility to outpatient health services and reasons for abandoning appointments with primary care physicians, as well as the staff resources of family medicine specialists.

Methods

The paper is based on the results of a nationwide statistical survey on the activities of medical facilities and the use of health care services in households. They provide information on the number of consultations provided in PHC departments and infirmaries by age group of outpatients, an assessment of the course of visits to the primary care physicians or main reasons for unmet needs for medical examination or treatment.

Results

The data collected show that in 2023, there was an average of almost five consultations per capita provided to outpatients in the PHC, and about 3% of household members did not use the services of a primary care physicians, despite the need.

Conclusions

The PHC is the most accessible form of health care, so it is important to emphasise the need for further investment in its development, especially in the context of the observed ageing population. The PHC provides the first line of contact between the patient and the health system, leading to early detection and treatment of diseases, resulting in improved population health, system efficiency and reduced medical expenditure in the long term.

Keywords

primary health care, health care system, PHC, health care statistics, an outpatient department and infirmaries of the primary health care

Objective

The aim of this paper, as well as the research project being carried out, is to present verification of the links between the characteristics of Internet use among Polish youth (the most popular content, dominant social media, time spent on the Internet, etc.) and subjective (declared, perceived) general health status. The cognitive impulse for the research project carried out in the panel perspective was the progression of broadband Internet in 2014, then – around 2016 and 2017 – the so-called smartphone revolution, and currently the expansion of AI mechanisms.

Methods

The project has been carried out for 10 years, last year a study was carried out verifying the potential impact of interactional strain resulting from the frequency of using the Internet on the decrease in the perceived state of health (N=315), the results of which were presented at the 4th KSP. This year, the study covered over 500 people (youth, young adults, purposeful selection), using correlation, determination mechanisms and a linear regression model, while maintaining the main research assumptions, adding to the tool a component regarding pornographic content and a sense of anxiety.

Results

In 2024, a statistically significant correlation was noted between the so-called ”social online mirror” and lowered self-esteem (r=-0.303). The meta-analysis of ESS 11 data (linear regression models) shows that high frequency of Internet use affects the decrease in subjective general health in 12% (r=-0.347, p?0.001) in Polish society and as much as 48% among young people (r=-0.694, r-squared=0.48, p?0.001). At the same time, interactional strain was found among young people resulting from a number of different network activities.

Conclusions

The research results are the basis for the implementation of broadly understood public policy for youth, as well as detailed preventive programs in terms of reducing the risk of addiction to all mechanisms of the network society that determine the reduction of subjective general health. The significance of the project is therefore purely ”applied” (implementative), and the post-research conclusions and their validity imply the construction of models of cross-sectoral activities to reduce the risk of long-term consequences of the virtualization of the lifestyle of contemporary youth.

Keywords

subjective general health, Internet use, Polish youth, Polish society, anxiety

Objective

This study examines the impact of perceived workplace unfairness on job satisfaction in Germany. In particular, we focus on differences between different groups with different pay levels. While previous studies have mainly focused on distributive justice, this study also considers procedural and interactional injustice, providing a multidimensional perspective.

Methods

Using data from the German Linked Personnel Panel (LPP) covering private sector companies from 2012 to 2021, we classify employees into wage quintiles to analyze differences in perceptions of fairness and job satisfaction. Multinomial logit models are used to estimate the impact of perceived unfairness on job satisfaction.

Results

Our results indicate that lower-wage employees perceive significantly greater unfairness in the workplace, especially with regard to pay, decision-making processes, and treatment by superiors. While job satisfaction increases with income, perceptions of fairness decrease. A downward trend in overall job satisfaction over time was also identified, suggesting that economic and organizational factors influence perceptions of fairness.

Conclusions

The results of the analyses reveal that perceived injustice has a stronger negative effect on job satisfaction than absolute pay, underlining the crucial importance of workplace fairness beyond pay alone. These results are consistent with previous studies on perceived fairness, but they challenge traditional economic theories that assume a direct link between income and well-being. Our study highlights the need for transparent pay structures, equal treatment, and inclusive decision-making processes to improve employee satisfaction.

Keywords

workplace fairness, job satisfaction, pay stratification, procedural justice, perceived injustice

Objective

The aim of the study was to develop and evaluate the U-smile method, designed to visualize and quantitatively assess the usefulness of new predictors in binary classification models, with particular emphasis on scenarios involving imbalanced data.

Methods

The U-smile method – comprising U-smile plots and the BA (Absolute Change), RB (Relative Change), and I (Net Reclassification Index) coefficients – was used to analyze the impact of new predictors in binary classification models. The study was conducted using real-world data from the Heart Disease dataset and simulated data, incorporating various class imbalance scenarios.

Results

The U-smile method effectively identifies and visualizes the contribution of new predictors to classification improvement, even under data imbalance conditions. The introduction of a three-level approach and the use of the I coefficient as a weighting factor in BA and RB plots enhance the interpretability of results.

Conclusions

The U-smile method provides a valuable tool for assessing the usefulness of new predictors in binary classification models, especially when data are imbalanced. Its application can support variable selection processes and improve the interpretability of predictive models across various fields, such as medicine or social sciences.

Keywords

binary classification, U-smile method, ROC curve, likelihood-ratio test, imbalanced data

Objective

The purpose of the presentation is to highlight the information asymmetries present in the Polish labor market and their potential impact on the difficulty of choosing a career development path.

Methods

To analyze the supply side of the labor market, a pilot survey was conducted among students of a technical high school in Rzeszów regarding their career choices. For the analysis of the demand side of the labor market, web scraping was applied to two job offer portals along with text mining techniques.

Results

The pilot study showed that, in most cases, young people do not know what profession they want to pursue in the future. They primarily acquire knowledge about the labor market on their own or through conversations with their parents. A comparison of job offer portals revealed that the image of the labor market varies significantly between platforms.

Conclusions

The Polish education system lacks a coordinated and well-thought-out system for guiding career development. The abundance and diversity of available labor market information do not make this task easier. A modern career counselor must adopt new techniques for collecting and integrating data in order to obtain a coherent and comprehensive picture of the labor market and to perform their profession reliably.

Keywords

labor market, career counseling, web scraping, text mining

Objective

The study aim to determine the influence of wolves on the actual decline in the cattle population in the Podkarpacie region in relation to other factors.

Methods

The research is conducted on the basis of tables of conflict situations kept by the communes in which conflicts between wolves and humans are most visible.

Results

Behaviour patterns of wolves are derived in the long-term, as well as the elements that significantly influence the worsening of the decline in the cattle population in Podkarpacie are determined.

Conclusions

Wolves are species that, when given the choice between obtaining food in the wild forests or obtaining animals kept on farms, chose the latter. There may be many reasons, including the lack of game that they hunt in the forests.

Keywords

wolf, Subcarpathia, ruminants, population decline

Objective

The aim of the presentation will be to indicate which territorial units play a key role in the commuting system.

Methods

The analysis will compare selected structural measures of graphs for the studies mentioned above, illustrating individual aspects of mobility. Selected R packages will be used and presented for loading data provided by public statistics, processing them and visualizing the obtained results.

Results

It is assumed that the obtained results will allow the identification of territorial units that are strongly connected to others in the commuting system, as well as those that act as connectors between different parts of the network. Comparing different centrality measures will make it possible to capture the different roles of territorial units in the spatial structure of commuting and highlight differences in their interpretation and analytical value.

Conclusions

Matrix data, crucial for such an analysis, are still rarely available in public statistics and should increasingly be the end product of other studies. The presentation will be based on data from the Population flows related to employment NSP 2021 survey.

Keywords

population mobility, graph theory, centrality measures, network analysis

Objective

The aim of the paper is to identify groups of provinces with a similar level of innovativeness, and the analysis of group membership changes in time.

Methods

Dynamic taxonomy (YT|Z) is a clustering of spatio-temporal data. Standardization is performed for the whole period of analysis. Then, the agglomerative Ward method and k-means were used for clustering. The results are shown in tables presenting the membership of identified groups, and how it has been changed over time, and which objects moved from one group to another. The average composite indicators for groups make it possible to say if the observed changes meant a general improvement of industrial enterprises innovativeness.

Results

The research deals with innovativeness of industrial enterprises in Poland’s provinces of NUTS 2 level, in 2009-2022. The innovation activity is characterized by 39 variables from 5 spheres: investments in innovations (2), public support (6), innovation activity (16), cooperation in innovations (4), effects of innovativeness (11). With dynamic taxonomy we are able to track the movements of provinces between groups and characterize the innovativeness level of dynamic groups.

Conclusions

Dynamic taxonomy made it possible to identify spatial changes of industrial enterprises innovativeness in Poland – statistical data published by Statistics Poland on innovation on regional level was not homogeneous for the 2009-2022 period. Multiple regression was found to be a good tool for missing data imputation.

Keywords

dynamic taxonomy, innovation activity, Poland’s provinces, industrial sector

Objective

The aim of the paper is a statistical analysis of the impact of the lagged unemployment rate and GDP growth rates on the unemployment rate in regions of the Czech Republic and Slovakia. In their study, the authors attempt to answer the following questions: Do changes in regional GDP significantly influence changes in unemployment rates at the regional level in the Czech Republic and Slovakia? Were the effects of the 2008 financial crisis and the COVID-19 pandemic noticeable in both countries? Did the previous unemployment rate in both countries have a significant impact on its subsequent level?

Methods

The study is based on descriptive statistics, unemployment rate growth equation, the Ordinary Least Squares (OLS) method, Generalized Method of Moments (GMM) and Fixed effects.

Results

Prague recorded the highest GDP level during the analyzed period. The capital also exhibited the lowest average unemployment rate. Bratislava stood out significantly from the other regions in terms of GDP. The lowest average unemployment rates were also recorded in Bratislava. In both the unemployment growth equation estimated using OLS and GMM (with some exceptions), for Czech and Slovak regions, the previous unemployment level had a statistically significant impact on its growth. The GDP growth rate in Slovak and Czech regions generally reduced the increases in the unemployment rate.

Conclusions

The results of the conducted analyses provide a solid foundation for formulating practical recommendations in the field of regional and economic policy. The conclusions drawn from the comparison of Czech and Slovak regions can serve as a basis for developing more precise tools to support regional labor markets, especially in areas where unemployment is of a structural nature. The methods used in the article can also be applied beyond the borders of the Czech Republic and Slovakia, as the described statistical and econometric techniques maintain their universality in a comparative context.

Keywords

descriptive statistics, econometrics, mathematical economics, regional development, unemployment

Objective

The Statistical Office in Lublin, as part of the implementation of tasks related to the acquisition of administrative data, has cooperated with the General Office of Building Control (GUNB) on the SOPAB system under development. The purpose of the presentation is to introduce SOPAB as a future source of construction data. The innovative solutions used in the newly-created registry will improve data quality. Considering the information resource of SOPAB, its creation will provide opportunities to obtain a wider set of data in the future than at present, and this in turn will create new opportunities for analytical and publication work.

Methods

No statistical methods will be used. The presentation is intended to introduce the SOPAB administrative system.

Results

Currently, consultation work is underway on the implementation of the assumptions of the system between the GUNB`s project task contractor, Poznan Institute of Technology, and the SO in Lublin. Their purpose is to ensure methodological consistency and completeness of the data to be obtained, necessary for official statistics to carry out the tasks included in the Statistical survey program of official statistics (PBSSP) and international obligations. SOPAB will contribute to reducing workload, eliminate the manual process of filling out statistical forms improving the quality and detail of the data collected, and streamline the process of obtai

Conclusions

In the future, the SOPAB system will become a source of administrative data in the field of building permits and tangible effects of construction. Ultimately, the expanded System will include solutions and tools for the complete development of results in the field of construction, the preparation of which will significantly reduce the workload and time consumption They will therefore serve both cooperating institutions and, consequently, the whole society.

Keywords

administrative data, construction statistics, digitalization, construction process

Objective

The aim of the presentation during the poster session is to present the position of the Polish scientific journal in the international scientific circulation. The history of the journal, citation indicators in international databases, special issues published in SiTns in recent years and the thematic scope of the quarterly will be presented.

Methods

Presentation of the history of the journal, thematic scope, Special Issues over the years during the poster session during the 5th Congress of Polish Statistics.

Results

Statistics in Transition new series is present in many databases, including Scopus. The journal aspires to join the elite Philadelphia list of scientific journals.

Conclusions

Statistics in Transition new series is a leading Polish scientific journal in the field of statistics. Its presence in many international databases, including primarily Scopus, indicates a very high substantive level of the journal.

Keywords

Statistics in Transition new series, scientific journals

Objective

A two-stage regression modelling for estimating quantile premiums in automobile insurance is considered. It combines logistic regression (LR) to estimate the probability of claim occurance with quantile regression (QR) to estimate the value of claim for single risk. Traditional approaches like Generalized Linear Models (GLMs) provide expected premium estimates but fail to account for extreme claims, leading to the need for quantile-based methods. Our aim is to study if ensemble techniques QR forest and QR boosting outperform the classical QRs, allowing the insurer to form profitable portfolio

Methods

The QR is used to estimate the quntile premiums in risk groups by minimizing the loss function. The model estimates the conditional quantile function, capturing heterogeneity in the data and addressing issues such as non-normality and heteroskedasticity. Then the ensemble methods: QR Forests and QR boosting is applied. This two methods combine the fexibility of machine learning with the interpretability of quantile regression. The QR forest model extends the random forest algorithm for the decision trees while the QR boosting employs gradient boosting techniques to iteratively improving weak

Results

This study highlights the importance of advanced quantile-based methods in estimating automobile insurance premiums, addressing the limitations of traditional models. The QR, together with QR forest and QR boosting, provides a robust approach for handling highdimensional data and accounting for extreme claims. Empirical analysis of the Polish automobile portfolio demonstrates that QR forest outperforms other methods in predictive accuracy, ensuring equitable risk distribution among policies.

Conclusions

The results emphasize the value of incorporating tail risk into premium estimation, leading to more conservative and reliable pricing strategies for high-risk scenarios. By integrating modern machine learning techniques, the research bridges the gap between traditional actuarial practices and contemporary data-driven approaches. Ultimately, the study show the way for more accurate and fair insurance ratemaking.

Keywords

quantile premium, automobile portfolio, quantile regression, logistic regression

Objective

The aim of the study is to develop a new price index based on transactional (scanner) data that satisfies both axiomatic requirements (in particular the identity and transitivity tests) and accounts for changes in product quality over time. The authors propose a new index – GEKS-GAQI (Geometric Asynchronous Quality Adjusted Index) – as an extension of the multilateral GEKS indices, which are widely used in inflation measurement based on scanner data.

Methods

An axiomatic and economic approach was applied in constructing the price index. The new GEKS-GAQI index is based on quality-adjusted unit prices and quantities, and uses a geometric averaging method of the sub-indices. The analysis was conducted using real scanner data on milk and coffee sales in Polish supermarkets, provided in the R package “PriceIndices.” In the empirical study, GEKS-GAQI was compared with other indices: GEKS, CCDI, TPD, GEKS-L, GEKS-AQI, GEKS-AQU, and GEKS-GL, both at the barcode (GTIN) level and the COICOP 6 level.

Results

The GEKS-GAQI index satisfies both the transitivity condition and the identity test, making it stand out among other multilateral indices. The analyses showed that the level of data aggregation significantly influences index values. At the COICOP 6 level, index values are more stable than at the GTIN level. Additionally, GEKS-GAQI produced values close to those of the GEKS-GL index, which also meets the identity test. Interesting differences between the multilateral indices under consideration were observed depending on the type of data (milk vs. coffee) and the level of aggregation.

Conclusions

The proposed GEKS-GAQI index provides a methodological and practical contribution to the development of modern inflation measurement methods using scanner data. It stands out from popular indices by combining quality adjustment features with formal properties, making it a promising option. Its use can improve the accuracy of inflation measurement and reduce distortions caused by the so-called chain drift effect. Moreover, GEKS-GAQI bridges the gap between the GEKS approach and the quality-adjusted Geary-Khamis method, opening new possibilities in the field of price statistics.

Keywords

Scanner data, multilateral indices, the GEKS-index

Objective

Development of econometric models for estimating market value and verification of the usefulness of such models in football club management

Methods

Econometric modeling was used. A linear model was proposed, the parameters of which were estimated using the LSM. In case the models showed heteroskedasticity of the random components, the generalized least squares (GLS) method was used to estimate the structural parameters. The subjects of the study were all footballers playing in the 2021 / 22 season in the top division in England (Premiership) and Poland (Ekstraklasa)

Results

A set of well-matched econometric models for valuing footballers playing in different positions (goalkeeper, defender, midfielder, striker) in the Premiership and Ekstraklasa was estimated. The research identified the most significant factors influencing the market value of players, and made it possible to identify overvalued and undervalued footballers compared to current market valuations conducted by the Transfermarkt portal

Conclusions

Overvalued and undervalued football players were identified. The proposed econometric models provide guidance to clubs as to which players are worth investing in and which are worth selling. Specific examples have been identified of footballers for whom a decision to buy / sell them based on the estimated model would lead to additional economic benefits. This is probably the first study indicating such a use of econometric modeling in football club management

Keywords

football player valuation, econometric modeling, football club management

Objective

The history of official statistics can be inquiried in at least three dimensions from a sociological point of view. Firstly, in institutional dimension of a development and reconfigurations of stiatistical institutions, organization of research, management and coordination of statistical undertakings. Secondly, in discoursive dimension of statistical discourse and its reshapings, statistics as a specialistic language with its own concepts, definitions, classifications, and nomenclatures, as well as typical styles, genres (e.g. scientific paper, analitical elaboration), their contexts (scienti

Methods

The presentation will try to grasp the history of statistics from a sociological point of view. The methodology will comprehend qualitative analysis of discourse, anthropology of writing, and source criticism.

Results

The application of methodology of the historical sociology and other social sciences to inquiry into the history of the official statistics will provide the means for understanding what visions and ideas were associated with the shifts in the statistical discourse.

Conclusions

The history of Polish statistics require an approach broader than pure factography. The presentation will try to answer the problem.

Keywords

official statistics, historical sociology, qualitative analysis, discourse

Objective

Nearly a 100 years ago, Poland’s distinguished anthropologist Jan Czekanowski indicated the importance of military conscription records regarding the conscripts’ height. In his view, these constituted a source of primary importance for anthropologists, sociologists and historians of the economy. Half a century later, these materials became the foundation of a methodological shift in the studies of the history of the economy focused on tracking the changes in the standard of living of broad masses of people in the past. The aim of this paper is to discuss the most important results of these st

Methods

The height, on the level of an individual, is determined genetically, but the genetic potential might yield different outcomes depending on individuals’ living conditions. If they are not optimal, the potential is not fully used. As a result, by studying larger populations and observing the pace and direction of changes in the mean stature, we can draw conclusions about the biological welfare of the population, social stratification, etc.

Results

Since 1965, studies have been conducted at approximately 10-year intervals on the basis of representative samples. Whether they are conducted or not depends on external funding and the willingness or its lack of institutions examining conscription-age young people. Historical studies are based on sources that have often been preserved randomly and by nature are not representative of the entire population. Nevertheless, they might serve as a valuable complement to contemporary research.

Conclusions

To conclude, it is necessary to return to Czekanowski`s postulate and make sure that official statistics collects aggregate data on height and body mass of the population of not only conscripts, but also primary and secondary school students. Even aggregate compilations of this type would constitute a valuable addition to data collected by Statistics Poland

Keywords

standard of living, examination of conscripts, examination of school students, height, weight, BMI

Objective

The aim of the paper is to attempt to answer the question about the relationship between age heaping and the ability to sign one’s name, along with their administrative context, on the basis of individual historical data drawn from marriage records dating back to the first half of the 19th century from two Kraków parishes. The history of statistics might view this problem as the transition from order-related and symbolic understanding of statistics by institutions to informational understanding.

Methods

The research into age rounding was conducted using Whipple’s and Myers’ indices. For each index, percentile confidence intervals were also calculated using the bootstrap method (R=1000). In determining the level of literacy, the study used the Maggiolo method, i.e. the assumption of a one-to-one correspondence between a person’s signature on the certificate and their ability to write. Marriage records for the period of 1810–1848 from of two Roman Catholic parishes in Krakow, the St. Florian and the St. Szczepan parishes, served as the basic source for this study.

Results

When individuals provided their age without any official verification, literate people did so much more accurately than those who could not write. The rounding of age could therefore serve as a good approximation of their numerical competencies and, more broadly, of human capital. When the age recorded in the sources was based on documents, however, the ability to write ceased to play any role in the accuracy of age reporting. This means that in the early 19th century, institutions were already able to legally enforce certain standards on the population that led to recording information with g

Conclusions

The analysis of the rounding of age in historical quantitative sources makes it possible to look at statistics not as just the collection of data, but as a social and institutional process, with all its errors, compromises and imperfections.

Keywords

age heaping, records of marital status, human capital, numeracy

Objective

The research problem focuses on the subject matter and thematic scope of Polish censuses in the 19th and 20th century (until 1988). The role of international models and standards in the construction of national censuses is described, as well as the importance of international statistical congresses, statistical agencies operating under the League of Nations (in the interwar period), the United Nations (in the second half of the 20th century) and the International Statistical Institute. The scope of the censuses and their role as a source of information for social and demographic research.

Methods

The study is based on a comparative analysis of international and national census documentation and forms from the 19th and 20th century. The subject matter and scope of the studied demographic and social issues, the continuity and discontinuity of certain groups of issues and the problem of the repetition of the scope of information were all reconstructed. In addition, the challenges associated with the organisation of censuses (personal censuses, administrative censuses, population registration based on various sources, etc.) were presented.

Results

Censuses began to be conducted in some European countries in the second half of the 18th century, and became a cyclical event in the 19th century. In the second half of the 19th century, censuses gained attributes of the universality, being conducted on a single day, directly and respondents being identified individually. International statistical institutions played an important role in the development of census procedures and standards. A set of common questions concerning demographic and social issues became one of the basic features of censuses in the 20th century.

Conclusions

Censuses constitute one of the most valuable sources for historical studies of demographics, levels of education, the living conditions of the population, its occupational structure and sources of income of societies. The thematic scope and set of census questions evolved under the influence of national and international standards, leading to the comparability of data in dynamic and spatial terms. Due to the fact that censuses were conducted at the level of the smallest administrative units, they also showed societies in different regional cross-sections.

Keywords

history of statistics, Polish censuses in the 19th and 20th century, historical demographics, statistical sources

Objective

The study aimed to assess the impact of ESG scores on the market risk of companies listed in the S&P 500 index. Specifically, it examined the influence of ESG performance on market risk both across the entire index and within selected sectors, including Financials, Health Care, Industrials and Information Technology.

Methods

In the first step, the GARCH-type models were fitted to the daily rates of return of 482 S&P 500 companies. Based on the GARCH-type models, we determined the dynamic Value-at-Risk (VaR) for each company. In the next step, the relationship between the VaR and the companies' ESG scores was identified for each day of the analyzed period. There were estimated dependence measures of 5%-VaR and 1%-VaR with companies' ESG scores. Considering that the relationship between VaR and companies' ESG scores may be non-linear, the analysis included not only the Pearson correlation coefficient, but also the S

Results

All dependence measures oscillate near zero during all the analyzed periods. The lower bound of confidence intervals is slightly above zero only in a few days in February 2022 and October 2023 – the beginning of the war in Ukraine and the beginning of the Israel-Hamas conflict. The effect is minimal and fleeting. Similar results were obtained for the companies from the Financials, Health Care, Industrials and Information Technology sectors. A more evident positive correlation between ESG scores and VaR is observed in the Industrial sector.

Conclusions

The conducted analysis confirmed a statistically significant correlation between Value-at-Risk and ESG scores exclusively for companies in the Industrials sector of the S&P 500 index. This finding suggests that, within this sector, lower market risk is meaningfully associated with higher ESG scores. One possible explanation is that companies in the Industrials sector may demonstrate greater commitment to environmental responsibility, social impact and corporate governance, which may be positively perceived by the market.

Keywords

GARCH models, ESG, S&P500 index

Objective

Safe-haven and hedging instruments are market assets that allow investors to protect their portfolios against losses. A key limitation of most existing research in this topic is the assumption that risk minimization should be the sole criterion in the evaluation of decisions relative to safe-haven and hedging allocation. This perspective fails to capture the real motivations of investors who perceive investments as a trade-off between risk and profit. In contrast to those studies, we propose to verify the hedge or safe-haven property in the prospect theory framework.

Methods

Four of the most often explored assets, i.e., Bitcoin, Ether, Tether and gold, are tested as hedges or safe havens against the G7 and BRICS stock market risk using the cumulative prospect theory framework. We estimate an optimal weight in a hedged portfolio of each candidate for a hedge or safe haven, as well as the effectiveness of the hedge or safe-haven strategy. We introduce the hedging effectiveness measure and show the impact of risk aversion and reference point selection on optimal hedging decisions.

Results

We found that Tether and gold can be useful assets in hedging or safe-haven roles. Bitcoin and Ether cannot increase weighted utility when included in a stock market portfolio. We also compare the results with the minimum variance approach, which is a benchmark for this research. We found that both methods similarly indicate the usefulness of individual assets in a hedge or safe-haven role.

Conclusions

Adopting a behavioral finance approach allows for a deeper understanding of the real motivations behind investors’ decisions regarding the selection of hedging and safe-haven instruments among cryptocurrencies and gold.

Keywords

cryptocurrencies, gold, hedge, safe haven, prospect theory

Objective

The aim of this paper is to examine the occurrence of ‘safe haven’ instruments in the context of alternative investments proposed in the literature.

Methods

We estimate the time-varying correlations between alternative investments and stock indices using the DCC-GARCH model. We consider time periods: 01.01.2018-31.01.2020, 03.02.2020-31.01.2022 (COVID-19 pandemic), and 01.02.2022-01.03.2024 (after Russian aggression in Ukraine).

Results

We can observe that in the period before the COVID-19 pandemic, natural gas was the dominant ‘safe haven’ instrument. During the COVID-19 pandemic, we could observe that gold, natural gas, CHF and Bitcoin were mostly ‘safe haven’ instruments. After the Russian aggression in Ukraine, we can observe that mainly the USD and the CHF were considered ‘safe haven’ instruments. During all the periods mentioned, we were able to observe changes in the ‘safe haven’ instruments.

Conclusions

The most important added value of this work in comparison to the existing literature is the inclusion of different countries (from America, Asia and Europe) for such different periods. We compared different situations in financial markets to show the occurrence of ‘safe haven’ instruments in various countries.

Keywords

safe haven instruments, gold, silver, Bitcoin, dynamic correlation

Objective

The aim of the paper is to present the participation of Polish statisticians in the process of creating a uniform methodology of statistical research. The foundations of official statistics were created during 10 Statistical Congresses in the years 1853-1885.

Methods

The presentation of the achievements of statisticians is based on an in-depth analysis of the literature on the subject and on the study of documents and source materials.

Results

Starting from the Second Congress, which took place in Paris, Polish statisticians, including August Cieszkowski (1814-1894) took an active part in the event. In 1885, the last congress in this form was held in London, and at the same time the International Statistical Institute was then established. On June 24, 1825, the statute was adopted and the authorities were elected. The first president was Sir W. Rawson. One of the founders of the Institute was Professor Tadeusz Piłat (1844-1923) – director of the National Statistical Office in Lviv. He was the chairman of the MIS committee.

Conclusions

The next ordinary member of the Institute was Józef Kleszczyński (1841-1900), a professor of statistics and administrative law, who in 1885 published in the journal 'Przegląd Polski' an extensive article entitled 'International Statistical Institute', which was the first work devoted to the Institute. In this article, Kleszczyński postulated the establishment of an international statistical institution of an official nature. This postulate was implemented only after World War I.

Keywords

International Statistical Institute, Polish statisticians creating European official statistics