Conference Paper BHPS-2005 Conference: the 2005 British Household Panel Survey Research Conference, 30 June -2 July 2005, Colchester, UK
Married or single: which shall I tick? Findings from a study of BHPS marital status data
We have been studying the union histories (including both marriage and cohabitation)
of persons aged 50 and over in Britain, using data from the BHPS, waves two to
twelve. This is part of a wider ESRC-funded project investigating social support in
later life. The aim with the union histories is to identify family disruption due to
divorce, separation, death or re-partnering.
We have found that the histories the reports of events before the
respondent's entry to the survey appear consistent with reported marital
status and household composition at the time they were collected, i.e. on the
household roster and individual questionnaire. (By marital status, we mean reported
legal marital status, actual marital status and cohabitation.) However, responses to
questions on marital status in succeeding waves show less consistency. This often
appears to be due to respondents' constantly changing assessments of their situation.
Using the variables xMLSTAT, xMASTAT, xMLCHNG and xSPPID, which appear
in every wave, it is apparent that reports differ from wave to wave without any
apparent change in the respondent's partnership status; for example, reported actual
marital status changes but the identified partner remains the same. Moreover, some
respondents have inconsistent values over these variables at the same wave.
Some of these anomalies can be explained by changing social conditions, including
the well-documented increase in cohabitation and the rise in legal or non-legal unions
where the partners do not co-reside. In addition, more older respondents than
expected did appear to have frequent changes of partner, which put their actual
marital status in a constant state of flux. But indications were also found that the
categories themselves are a problem to respondents, for example: what constitutes a
legal separation? Can a couple cohabit without coresiding? This finding has
implications for the interpretation of union status reports in cross-sectional surveys
(e.g. the General Household Survey) and indeed in the Census.