Application of the Multinomial Logistic Regression Model to Analyze the Impact of E-Government on Youth in Albania

E-Government is using digital technology to improve services and increase civic participation. By digitalizing public services and transforming e-governance, governments are changing the interaction between people and institutions, to make operations more transparent, efficient, and accessible. More important is the fact that digitalized transactions reduce the burden on citizens and businesses by making public services faster, more convenient, and less costly. Understanding the e-government level of use and the experience of young people is important for successfully implementing public service digitization. This study aims to analyze the various factors that affect young users' experience of e-government and how the digitalization of public services impacts Albanian youth, focusing on their perceptions, experiences, and how these services influence their integration into the digital world. Employing a multinomial logistic regression analysis through a survey of young people from different parts of the country, the objective is to understand how they feel about the e-governance experience. The study explores the relationship between user experience (binary dependent variable) and key independent variables, including ease of use, data security, cost reduction, time reduction, bureaucracy, human error reduction, system malfunctions, lack of skills, and access to the internet. To identify the most relevant predictors for the multinomial outcome variable, we conducted a stepwise selection process using multinomial logistic regression which considers all possible variables in the reduced dataset from the survey. To ensure the robustness of our selected model, we performed 10-fold cross-validation using the `cv.glm` function in R. Variance Inflation Factor (VIF) scores were calculated to evaluate multicollinearity among predictors. Results: The logistic regression model reveals that Corruption and system malfunctions significantly influence users’ experience, of young people and are more likely to report satisfaction. In the final step of the stepwise model selection, only two variables were found to be important and with a low VIF: Corruption and System Malfunction. Corruption and system malfunctions emerge as strong predictors of experience satisfaction with e-government. The model demonstrates good fit and predictive ability, as indicated by statistical tests and graphical analyses, including a Predicted Probability Plot. The findings show that many young people view e-governance as a positive step towards modernizing public services and promoting transparency. However, there are still challenges, like ensuring access to technology and improving digital literacy. Addressing these issues will help maximize the benefits of digital transformation for all young citizens.