The Demographic Transition in Southern Africa: Yet Another Look at the Evidence from Botswana and Zimbabwe

Abstract

The findings of Thomas and Muvandi about overstatement of fertility decline in Botswana and Zimbabwe are questionable when appropriate methodology is employed. The reexamination of evidence on fertility decline in Botswana and Zimbabwe shows that educational distributions in 1984 and 1988 did not differ statistically and significantly and that the mean number of children ever born was by standardizing the educational distribution in each survey. Period fertility rates are a better sense of assessing period fertility declines and adjustment of educational distributions would not appreciably affect the magnitude of fertility rate decline. Topics in contention between the two studies are questionnaire wording changes in the educational system over the time (school starting age and requirements for completion of primary and secondary school real increases in education smaller shifts in the large group of women with some primary education data processing rules and sample similarity. In the Thomas and Muvandis comparisons of children ever born there were 34 out of 36 comparisons within education groups for the same cohort of women in 1984 and 1988 that were within two standard errors of each other which supplies overwhelming evidence that the numbers of children ever born for the same cohort of women in the same educational groups are in fact the same. The Thomas and Muvandi comparisons of women aged 25-34 and 35-44 years standardized for educational distribution differences were affected by sample selection bias differential reporting of education with no sample selection bias and real changes in education between cohorts. The results of comparing the same birth cohort indicate that there was little difference in the reports of the 1950-59 birth cohorts from the two surveys. A comparison of total fertility rates standardized by education show that Contraceptive Prevalence Survey (CPS) Demographic and Health Surveys (DHS) data (.13 child) and 1.55 children less for CPS data in Zimbabwe; in Botswana CPS declines were 12% and DHS 7% which was interpreted as evidence of limited impact from selectivity bias. The 1994 DHS should provide ample evidence of Botswanas and Zimbabwes unequivocal fertility decline.