Intercensal fertility estimation in Kenya

Abstract

This study examines indirect techniques of estimating intercensal fertility levels in Kenya. Parameters,Crude birth Birth Rate, Total Fertility Rate, Net Reproduction Rate, Gross Reproduction Rate and the Mean Age at Birth for the 1969 to 1979 intercensal period are estimated at National, provincial and district levels. These are estimated by applying several standard indirect procedures and models of estimating fertility which are also compared. To obtain estimates that reflect effects on a hypothetical cohort of a prolonged exposure to vital rates in operation during the 1969 to 1979 intercensal period, the standard procedures are adjusted by the additive synthetic adjustment technlque. The age-specific growth rate adjustment technique, functioning as both an indirect procedure and an adjustment technique, is also used to estimate fertility parameters that reflect exposure of a hypothetical age cohort to past rates of attrition and entry into the population. IhlS also reflects, among other things, the effect of vital rates and migration rates prevailing during the 1969 to 1979 intercensal period. The techniques have also been applied in evaluation of data quality. Chapter One introduces the study by stating the problem, objectives of the study, significance of the study, literature on historical and technical backgrouds of techniques, sources and quality of data and definitions of parameters. Chapter Two discusses P!F ratio methods, namely the Brass and Coale- Trussell P!F ratio methods. Chapter Three examines techniques

that base estimations on only reported lifetime fertility, 2 namely the Coale-Demeny P3 /P2 formula, and Brass-Rachad's 4 P2(P4!P3) formula. Chapter Four Describes and discusses the relational Gompertz model. Chapter Five discusses the agespecific growth rate adjustment technique. Chapter Six gives a summary of the findings, conclussions and recommendations. This is followed by references/bibliography and appendices. National, provincial and district level analysis shows that indirect techniques that estimate intercensal fertility levels in the presence of trends donot perform as well as those that rely only on lifetime fertility data. Also evident is the disparity between data on births reported in the twelve months peceeding each census and data reported by the same women regarding the number of children ever born. Synthesising data to refer to hypothetical cohorts theoreticaly minimises the effects of intercensal changes in vital rates and patterns of error on fertility estimates for the same period. But, still, additive synthetic adjustment technique amplifies the margin of error created by such changes. Also, the age-specific growth rate adjustment technique is reliable only if changes in the composition of age cohorts are adequately adjusted for. The latter is especially sensitive to effects of migration.