Gender Attitudes to Gender Roles in Britain between 1991 and 2007

Declining participation rates are a major challenge for longitudinal surveys, not only because this results in progressively smaller sample sizes but also because it increases the risk of the sample becoming less representative of the population of interest. Consequently there is significant motivation to improve our understanding of how a respondent‘s survey experience in an earlier wave 19 relates to their likelihood of future cooperation. Existing evidence is limited, particularly for face-toface surveys and within the UK context. Some studies suggest that respondents‘ concerns at initial contact and attitudes towards the survey are strongly correlated with future response (McCarthy & Jacob, 2009; Branden, Gritz & Pergamit, 1995). It has also been suggested that item non-response could indicate lack of motivation which could be related to future nonresponse. Others have examined the relationship between the length of the previous interview and response at the subsequent wave, with inconsistent findings. If anything, the relationship between interview length and future response has been found to be a positive one, with longer interview inferred to be yet another indicator of higher level of cooperation and engagement – therefore higher chances of future participation too – by the respondent (Branden, Gritz & Pergamit, 1995). In this paper we will examine the relationship between available indicators of the ‗survey experience‘ on Wave 1 of Understanding Society and subsequent response on Wave 2. The indicators considered will include initial level of household resistance and queries on the doorstep, the number of calls and issues required to acquire a productive interview, the interviewer observations about the respondents‘ overall level of co-operation and suspicion across all interviewed adults in a given household, and Wave 1 item non-response. We will also consider the overall interview burden, looking the individual interview as well as other parts of the Understanding Society survey instrument including the paper self-completion questionnaire, household-level enumeration and household questionnaire, and proxy interviews. We will conclude by suggesting some practical steps to identify households most at risk of non-response and develop contact strategies that could be used to mitigate that risk.