The effect of sex composition on contraceptive use

This study used the 1998 Kenya Demographic and Health Survey (KDHS) data to establish the effect of sex composition on contraceptive use. Specifically, the study sought to: Determine women's preferred sex composition of their children; Determine the association between sex composition of living children and contraceptive use; and make policy recommendations. Results yielded by cross tabulations and the chi-square test (X2) showed that there was a significant association between sex composition of living children and contraceptive use. Precisely, the X2 value = 126.635 at 20 degrees of freedom and a=O.Ol. This resulted to the rejection of the null hypothesis, Ho, stating, there is no significant association between sex composition of living children and contraceptive use. By implication, the alternative hypothesis, HI, was accepted. That is, there is a significant association between sex composition of living children and contraceptive use. Arnold's model was used to determine the existing sex preference and its effect on contraceptive use. The assumption of the model is that women who are most satisfied with the sex compositions of their living children will most likely use contraceptives. Overall, an existing preference by women to have a balanced number of boys and girls was reflected. Besides, there was a slight preference for sons over daughters. Results indicated that in absence of sex composition preference, contraceptive use would increase by approximately 6.1 percent. Further analysis showed that regions were coupled with sex preference variations. Central, Coast, Western and Rift Valley provinces were characterized by sex preferences of at least one child of each sex although the preference was slight in Central province. Apart from Central, the other three provinces had moderate preferences for boys over girls. Eastern province on the other hand was marked by a desire for a balanced number of boys and girls as well as a mild preference for boys over girls. As for Nyanza province, strong sex preferences of boys over girls were witnessed. In all the regions, absence of sex preferences would have seen an increase in the proportion of women who did not want more children by a range of between 7.7 - 22.4 percent. Educational variations in sex preference and the effect on contraceptive use were also considered. Sex preferences witnessed among women who reported to have had no educational attainment as well as those with primary education were characterized by a preference for at least one child of each sex. Son preference was also visible though moderate. Contraceptive use in absence of the sex composition preferences would have risen by 8.1 and 8.4 percent for those with no education and primary education respectively. Among secondary and higher education holders, a preference for an equal number of girls and boys was observed in tandem with a slight son preference. Within this category, the highest contraceptors were those with a balanced or a near balance number of boys and girls. Although there were traces of son preference, the differences in one-sex compositions within a parity were not large. In absence of these preferences, contraceptive use would have risen by 4.7 percentage points. Recommendations made were categorized into two: (i) Policy recommendations and recommendations for further research. For the former, educational campaigns should be introduced to encourage couples to be satisfied with the sex composition of their children. Besides, women's education should be promoted since it helps them counter with social and cultural values that may discriminate against use of contraceptives. On the other hand, recommendations for further research were: Replicating the study in an urban situation so as to determine the existing sex preferences and the effect on contraceptive use; carrying out a similar study among men to determine their sex preferences and the effect on contraceptive use; conducting a qualitative research to establish the causes of sex preferences and the reasons for the variations; and to test the significance of the differences between actual contraceptive use and use in absence of sex preference.