Non-standard employment modeling in the Republic of Kazakhstan

Abstract

Purpose: To determine the factors affecting non-standard employment in the Republic of Kazakhstan and their quantitative characteristics using the SmartPLS software.
Methods: Over the course of research we have been using the methods of sociological survey, structural equations modeling (SEM) based on SmartPLS, and partial least squares (PLS).
Results: We have hypothesized and tested the influence of such factors as education, social security, the human factor and digitalization on non-standard employment. We have assessed all test tasks and the suitability of the test entity. Using the Cronbach's Alpha coefficient, we have tested the internal consistency of the test questions and measured the effect of each question on the latent variable. Most of the indicators have high indicators, with the exception of the “Digitalization” factor. The low value of this factor is justified by the heterogeneity of the test responses. We have calculated average variance extraction (AVE) and reliability (Composite Reliability) coefficients of the model. We have checked the model variables for multicollinearity and calculated the determination coefficient.
Conclusions: Results of the analysis show that today the major issue is the lack or low accumulation of human capital among non-standard employees. The value of the R-square determination coefficient for the dependent variable “Non-Standard Employment” has a high value (0.75), which indicates that the factors included in the model describe well and have a high degree of influence on it. In general, the structural analysis has shown that the resulting model is adequate and built fairly well. Path Coefficients, reliability and validity coefficients are high enough to assess and analyze non-standard employment.