THE SOCIOLOGICAL POTENTIAL LONGITUDINAL SURVEYS IN NEW ZEALAND: EDITORIAL COMMENTS

While experimental studies are seen by some as the ‘golden standard’ for social research methodology these often lack a linkage to the real-world and their findings lack generalisability as a result. Instead, longitudinal studies capture most of the advantages of experiments while capturing real-world material and are a combination of cross-sectional and time-series. New Zealand is host to at least two long standing longitudinal studies of international importance, while the recent decade has seen a major set of developments of new projects. This editorial brings attention to the importance of these in general, and especially for New Zealand sociology.That potential has yet to be immediately garnered if publications in New Zealand Sociology are any gauge. Although several updates of the FWWP panel project have been published (eg. von Randow & Crothers, 2014) the only article directly using longitudinal data is a commissioned one appearing in a special issue on Social Class. New Zealand sociologists seldom even refer to the data produced. However David Fergusson from the Christchurch project publishes regularly in sociology journals and hence is undoubtedly New Zealand’s most well published sociologist, and other longitudinal projects have also published directly sociology-relevant articles. Pearson & Thorns (1984) claim their historical time series study is longitudinal but I think they’re using the term in a different sense.Beyond longitudinal and panel studies there are other combinations and mixes. Cohort studies may add in fresh waves of recruits while remaining largely cohort studies whereas others may retain a cohort component while largely being panel studies. And after all, many studies have ‘before and after’ phases without attention to long term changes (and so are excluded from treatment here). There are also ongoing survey research series without a longitudinal component: GSS, ISSP and NZVS, the University of Auckland national youth health & wellbeing studies. Administrative data often has a longitudinal component. (Examples are the Ministry of Social Development Benefit Dynamics dataset and the Linked Employer-Employee Data (LEED) more recently amalgamated into the IDI). Thesis work may involve some overtime work, although unless they are working within a larger project, thesis completion times would usually preclude mounting their own longitudinal studies.The (then) Families Commission published a review of New Zealand longitudinal studies in terms of their relevance to the research work of the Commission and thus through a ‘family lens’ (Poland & Legge, 2005) and Mckenzie and Carter (2010) has also published on this topic, exploring the fit between longitudinal studies’ definitions of family and the concept of ‘whanau’. This editorial extends and updates these two reports. I provide a top-level round-up and cite only the most crucial publications as general information is mainly carried by websites: so only some books of findings or methodological accounts are referred to here. Finding out about some of these studies was difficult and there seemed little on their websites, so there is some ‘missing information’. Further more detailed work is warranted.General Methodological PointsLongitudinal surveys follow the same group of respondents over time (sometimes over long periods of time such as decades) with repeated collection of the same variables. There may be disciplinary differences in focus. Sociology uses longitudinal surveys to study life events throughout lifetimes or generations, whereas psychologists study developmental trends across the life span, and medical researchers attempt to pinpoint antecedents (and consequences) of the advent and course of diseases. Examination of turnover (or churn) between surveys rounds allows estimation of durations. Because of the repeated observation at the individual level, they have more power than crosssectional observational studies, by virtue of being able to exclude time-invariant unobserved individual differences, and by virtue of observing the temporal order of events.