Qualitative perceptions of the meaning of “headship” and female-headed households in post-apartheid South Africa

Abstract

The identification of one household member as the head of the household remains a feature of household surveys conducted by Statistics South Africa. While the analytical relevance of this practice has been critiqued and while many national statistics agencies have abandoned the concept of a household head altogether, researchers in South Africa often use the characteristics of the household head in order to classify households. In particular, recent research has documented a rise in female headship in South Africa and a growing gap in poverty risks between female- and male-headed households in the post-apartheid period. Some of this work has also shown that the increase in female headship is due to the growing incidence of women living in households without men. The way that headship is assigned and what it actually means, however, is something of a “black box” in social science research. This paper presents the findings from a qualitative investigation of headship in South African households. The results suggest that most respondents attach meaning to the notion of a household head but that, as expected, some clear contradictions in the way that headship is assigned were encountered in the data. The paper concludes with a discussion of the implications for using the characteristics of the head (and gender in particular) as a way to classify households and identifies some lessons for survey research protocols.

Keywords:

• headship
• households
• poverty

Acknowledgements

The author would like to thank Dorrit Posel and the National Research Foundation’s South African Research Chairs Initiative (SARCHI) for support in the completion of the study. Thanks must also be extended to Kelly Hallman and the Population Council for permission to re-interview participants from an earlier study. The author is also indebted to Thembekile Zulu and Nokwazi Thusi for their important contributions to the fieldwork and translations. Finally, the author would like to thank the two anonymous reviewers for their detailed suggestions and comments.

Notes

1. In some countries, the use of headship in censuses and household surveys has been abandoned altogether in favour of other reference categories. In the USA and Canada, for example, the census no longer asks respondents to identify the head of the household due to concerns with ambiguity around the term “head” (Haughton and Khandker 2009). Indeed, it seems as though the use of “headship” is largely out of use in many developed countries (Ziehl 2001). In some contexts, a number of terms have been used in place of “female-headed” to describe, in greater detail, the type of household structure and its relationship to gender. Some of the more common terms include: “female-maintained,” “women-maintained,” “female-led,” “mother-centered,” “single-parent” or “male-absent” (see Buvinić and Gupta 1997).

2. The only prompt given in the questionnaire was that the head of the household is the person with the “primary decision-making responsibility.” Otherwise, respondents used only their own judgement to identify the heads of their respective households.

3. All of this information was available a priori through the ongoing longitudinal study.

4. Statistics South Africa’s instructions to fieldworkers on how to identify the household respondent typically suggest that this person should be any “responsible adult” who is able to provide information on other household members.

5. The rationale for this approach is that a follow-up question probed whether headship has any meaning to respondents. In the pilot interviews, we found that respondents would refer back to the definition provided by fieldworkers so the roster section of the interview schedule was revised to remove all prompts on how to assign headship.

6. This term translates directly to “the household,” but respondents themselves used it interchangeably with the term ekhaya which means “the home.” Similarly, the respondents themselves often preferred the term inhloko which translates directly to “the head.” During the pilot interviews, both inhloko and oyinhloko were used by the fieldworkers and respondents themselves indicated that they had similar meanings in the context of the discussion of household headship. In the interest of consistency, the term oyinhloko was used in the interview schedule, even though respondents often used the term inhloko.

7. The respondent from the ongoing study was a young person in each household who is participating in a life skills programme. The respondents in the present study were adults (18 and older) who were able to speak knowledgeably about the circumstances of their household (i.e. the same criteria that are often applied to censuses and household surveys). Approximately, one year elapsed between the original survey and this qualitative study.

8. As with Statistics South Africa’s questionnaires, the household roster presented to respondents did not contain a clear definition of either the household or the head.