14 Appendix C - The Kenya DHS Data

14.1 Kenya DHS

The 2022 Kenya Demographic and Health Survey (2022 KDHS) was conducted by the Kenya National Bureau of Statistics (KNBS) working with the Ministry of Health (MoH), in the first half of 2022. The Demographic and Health Surveys (DHS) Program, funded by the United States Agency for International Development (USAID) and operated by ICF, provided technical assistance.

The survey selected a nationally-representative sample through a complex stratified sample design. The stratification layer was whether the enumeration area was urban or rural). For inferring population statistics and parameters using the data, it should be weighted according to the sampling design. Because the data is used in this book for conducting introductory, exploratory research, we do not use sample weights in any of the analysis. Readers should be aware of this limitation!

Reports and survey data are publicly available (KNBS and ICF, 2023; KNBS and ICF, 2024). The survey sampled 42,300 households in 1,692 enumeration areas (clusters or villages) randomly selected. In each household, all women aged 15–49 were eligible to be interviewed, and in half of the sampled households, all men aged 15–54 were eligible to be interviewed. Women and men had to be “regular” members of the household or had to have slept in the household the previous night.

Earnings were reported, by these adult respondents, as monthly earnings. In-kind earnings were estimated, by the respondent, as a cash value. Responses were, of course, in Kenyan shillings (KES). The summary report for the 2022 Kenya DHS notes that there were responses from 14,983 women and 10,387 men, resulting in the following average earnings (KNBS and ICF, 2023:p. 498).

“Table 15.1.3 shows the average monthly earnings in the last one month before the survey for all employed women and men who were paid in cash or kind for their work, irrespective of their marital status. Those who were paid in kind were asked to provide the value of the amount received in kind. Average earning is calculated excluding respondents who did not work in last month or did not know how much they earned for their work. Average earnings for women (KES 12,166.9) are lower than for men (KES 18,594.9). Women report lower average earnings across all age cohorts relative to males.”

14.2 Variables used in the book

We used the data files KEIR8CFL.DTA and KEMR8CFL.DTA, the women’s and men’s individual records, and adapted them and limited the number of variables, to create the dataset used in this book.

Years of education

We use the years of education variable textttv133 for women and mv133 for men. The variable is standard in DHS surveys, often reported as “education in single years.” There were no missing values for this variable for the dataset of adult men and women.

Earnings

We recoded the monthly earnings variable s916a for women and sm606a for men to take into account missing values. We retained only observations (for adult men and women) who reported non-missing values. The largest source of missing values was, of course, due to people who did not do remunerated work (whether paid in cash or kind) outside of housework. This is usually more often the case for women, many of whom reported they only did housework), but some men also responded that they do not do work outside of housework.

Selecting only respondents who worked outside the home resulted in a sample of 9,725 men and 12,361 women, for a total of 22,086 observations. While there were more women initially interviewed (men were only interviewed in half of the households), the age range for men was slightly larger, and more women than men did not work outside of the household.

We converted the values into United States dollars using the 2022 PPP exchange rate of 43 shillings per dollar provided by the International Monetary Fund. The source of the PPP exchange rate used by the IMF is explained in the website IMF (2024):

“The Purchasing-power-parity (PPP) exchange rate (or conversion rate) between two countries is the rate at which the currency of one country needs to be converted into that of a second country to ensure that a given amount of the first country’s currency will purchase the same volume of goods and services in the second country as it does in the first. In the WEO online database, the implied PPP conversion rate is expressed as national currency per current international dollar. The International Comparisons Program (ICP) is a global statistical initiative that produces internationally comparable Purchasing Power Parity (PPP) estimates. The PPP estimates maintained and published by the World Bank, the OECD, and other international organizations, are used by WEO to calculate its own PPP weight time series. Currently, WEO PPP exchange rates are based on the ICP’s 2017 report.”

Other variables

We selected a limited number of other variables, in addition to earnings, gender, age, whether urban or rural, and years of education, to include as possible covariates in regression analysis. We included three dummy variables for religious affiliation (Christian mainstream Catholic or Protestant, Christian evangelical, and Muslim). Together these account for about 95% of respondents. We included one ethnicity variable, for whether the respondent identified as a member of the Kikuyu ethnic group, long one of the politically important ethnic groups in Kenya, and residing primarily in central Kenya around the capital Nairobi. The unweighted proportion of respondents who identified as Kikuyu was 17%. Other variables included were years lived elsewhere (other than the enumeration area where the survey was conducted), wealth group (integer from 1 to 5, indicating the quintile of the wealth of the household, as determined by an inventory of household assets), and month of interview.

14.3 References

IMF (2024). World Economic Outlook (WEO) Frequently Asked Questions. ENG.

KNBS and ICF (2023). Kenya Demographic and Health Survey 2022: Volume 1. Tech. rep. Nairobi, Kenya, and Rockville, Maryland, USA: KNBS and ICF.

KNBS and ICF (2024). Kenya Demographic and Health Survey 2022: Erratum report. Tech. rep. Nairobi, Kenya, and Rockville, Maryland, USA: KNBS and ICF.