Abstract

Survey-based estimates of electoral non-participation in Africa are often affected by social-desirability bias, measurement error, and inconsistencies in the operational definition of voter apathy across studies. This study proposes a multidimensional statistical measurement framework for decomposing voter apathy into three latent constructs: attitudinal disillusionment, motivational withdrawal, and behavioural abstention. Using a purpose-designed cross-national survey conducted in Nigeria, the study applies list experiments and randomised-response techniques to reduce response bias and obtain bias-adjusted prevalence estimates. The statistical analysis incorporates descriptive statistics, survey-weighted prevalence estimation, design-based variance estimation, and regression modelling to examine variations in voter apathy across demographic, socioeconomic, and institutional characteristics. Confirmatory factor analysis (CFA) and item response theory (IRT) models are employed to evaluate construct validity, item discrimination, measurement precision, and the underlying dimensional structure of voter apathy. Furthermore, multigroup confirmatory factor analysis and measurement-invariance testing are used to assess the comparability of the proposed measurement instrument across countries and cultural contexts. Where appropriate, latent-variable regression models are applied to identify factors associated with the different dimensions of voter apathy. The study aims to develop a statistically validated, multidimensional measurement instrument that supports reliable estimation, cross-national comparison, and evidence-based electoral participation research in Nigeria.