Abstract: Purpose: The objective of the study is to examine whether a wood processing company size affects the differentiation of workplace safety hazards as well as to investigate the influence of features characterizing an occupational accident casualty on their injury severity, considering the company size. Methodology: The study used non-aggregated data obtained from the Central Statistical Office, Poland. The data for analyzes were prepared through quality diagnosis, cleaning and transformation. Variables of no informative value were excluded from further investigation. Statistical tests were performed implicating the need for independent analyzes for two data subsets referring to: micro and small enterprises (employing up to 49 persons), and medium and large enterprises (employing 50 persons or more). For each of the two company groups, a logistic model was developed classifying the occupational accident casualty injury severity based on the casualty characteristics. In each case, the classification quality was assessed using a test data set. Findings: It was shown that the enterprise size had an impact on the severity of accidents at work and that the proposed method of classifying enterprises by size into two categories was justified. Explanatory variables in logistic models were interpreted according to their importance and intensity of influence on the explained variable. Practical implications: The obtained results can be used in the development of materials on occupational safety risks for entrepreneurs and OSH services. Social implications: Each type of economic activity carries various risks. Occupational accidents pose a serious social and economic problem. Research in the field of occupational safety allows a better understanding of the nature of such accidents and makes it possible to take effective preventive actions which, however, can depend on a company size. Originality/value: On the basis of the obtained results, it is possible to identify the factors influencing the severity of occupational accidents in wood processing companies according to their size. The research also showed that bivariate multiple logistic regression is an appropriate tool for analyzing occupational accident data.
Pełny tekst 
DOI Spis treści