Geoadditive modelling of adolescent age of sexual initiation among Nigerian women

Abstract

This study investigated the geographical variations in adolescent age at sexual initiation among Nigerian women aged 15-49 years using geoadditive binary probit model. The effect of current age was assumed to be nonlinear. Geographical information at state level was also explored in the model to assess the impact of spatial distribution of adolescent sexual initiation. Effects of all categorical covariates were assumed to be linear and hence estimated in a usual parametric form. Inferences were based on Markov Chain Monte Carlo (MCMC) techniques, while model selection was based on Deviance Information Criteria (DIC). Findings revealed significant spatial variations on level of adolescent sexual initiation among Nigerian women as well as a clear evidence of nonlinear effect of age.

Keywords: Adolescent, Geoadditive, Nonparametric, Spatial-effects, Bayesian