NON-LINEAR DECOMPOSITION OF CHANGES IN SEXUAL BEHAVIOUR AMONG YOUTH IN FOUR SUB-SAHARAN AFRICA COUNTRIES, 2004-2018

Abstract

Sexual behaviour functions primarily to assure reproduction, and sexual activity also contribute to HIV and sexually transmitted infections. The quest to prevent HIV, STIs, and unplanned pregnancy among youth in Sub-Saharan Africa make it necessary to regularly assess if there are changes in sexual behaviour. This study aims to investigate factors influence changes in sexual behaviour among youth in four Sub-Saharan African countries. Demographic Health Survey (DHS) data for two-time points with a minimum of 10years interval for four selected countries in Sub-Saharan Africa between the year 2004 to 2018 was used for this study. Chad, Uganda, Lesotho, and Nigeria were the selected countries with a total sample size of 74,138 for the youths age 15-24years age. Analysis for males and females were done separately, and the percentage change in five sexual behaviour indicators and predictors were determined. The preceding and recent surveys were merged for each country and gender before the non-linear decomposition technique was applied. The non-linear decomposition technique was used to assess the gap between the changes in characteristics effect and the contribution of socio-economic characteristics, culture/environment, media exposure to the gap for two sexual behaviour indicators with the highest magnitude of change used in each country. The result showed that each country and gender have different percentage changes in sexual indicators, with premarital sex having the highest magnitude of changes for both males (115.5%, 26.9%,72.8%,46.1%) and female (26.4%,44.4%,67.4%,30.4%) in all the four selected countries (Chad, Uganda, Lesotho, and Nigeria). Lesotho's non-linear decomposition models showed a decrease of 11.4% and 6.1% in multiple sexual partners and premarital sex among the male youths between 2004 and 2014. Socio-economic characteristics made the largest contribution to reducing the two indices of sexual behaviour, explaining 54% and 90% of the endowment effect. Also, there was a substantial increase of 29.5% in multiple sexual partners and a decrease of 32.5% in premarital sex among the female youth between the two surveys. Socio-economic made the largest contribution to the effect, explaining 57% and 194% of the endowment effect for multiple sexual partners and premarital sex, respectively. But for Nigeria, there was an increase of 10.5% in multiple sexual partners among male youths between 2008 and 2018. Media exposure made the largest contribution to this rise, explaining 444% of the endowment effect. For premarital sex, there was a substantial decrease of 23.1% between the surveys. Culture /environment made the largest contribution to the decline, explaining 135% of the endowment effect. A similar pattern was observed among the female youths. Non-linear decomposition was able to identify and explain the separate contribution of group differences in assessable characteristics. The analysis showed changes in sexual behaviour among the youths in four selected countries, and those changes were explained by changes in sexual indices across the countries and the explanatory factors.