Presentation on theme: "Mixed methods in longitudinal research Nick Buck UK Longitudinal Studies Centre University of Essex."— Presentation transcript:
Mixed methods in longitudinal research Nick Buck UK Longitudinal Studies Centre University of Essex
Overview This presentation comes from quantitative survey perspective, But main aim not to ask how qualitative methods could supplement research with quantitative motivation Rather the aim is to consider in more general terms the rationales for longitudinal research, and consider how far they should motivate mixed methods research This may involve use of language from quantitative research which is considered inappropriate in qualitative context, but perhaps there are analogues? Distinguish full and partial longitudinal mixed methods, on basis of whether use longitudinal approach with all methods or only one
Why longitudinal research? Longitudinal research defined here as using information collected from the same subjects at more than one point in time (i.e. not including retrospective life-history research based on single interview) LR is normally more expensive, complex and time consuming than research based on single observations, both in data collection and analysis stages (if separable) So requires special justification At its simplest, rationale is requirement for information on change or stability at the level of the individual subject (NB not necessarily social change)
Analysis of individual level change In quantitative LR concern is with change/stability of circumstances, behaviour, attitudes, values etc. It is recognized that there is a measurement problem because of distance between survey response and real behaviour etc (and also theoretical construct). Issues of reliability and validity are more salient in LR because more obvious than in cross-sectional research. For full longitudinal mixed methods need to raise the question whether there are analogues for these issues in qualitative LR, and what where the focus on change/stability lies in qualitative LR.
Quantitative LR focuses on: 1.Phenomena which are themselves inherently longitudinal, (eg poverty dynamics, employment instability, social mobility, social attitude stability); 2.Investigating causal processes, (e.g. impacts of unemployment on mental health or of child poverty on later life chances); 3.Controlling for the effects of unmeasured fixed differences between subjects; 4.The study of social change, separating out age, period and cohort effects; 5.Establishing the effect of a treatment in an experimental or quasi-experimental design, or comparing periods before and after the introduction of public policy.
Scope for mixed methods 1.Inherently longitudinal phenomena – clear scope for full and partial methods 2.Causal processes: partial methods? Understanding process in more detail 3.Controlling for unobservables: limited scope? Plausibility of assumptions? 4.Age, period and cohort: relies on large scale, long duration studies therefore limited scope? 5.Policy effects: similar to 1 ? 6.Other rationale coming from qualitative LR?
Types of longitudinal mixed methods: partial approaches Longitudinal survey as sampling frame for qualitative study (selection of subjects on basis of life history) LS as context for qualitative study (generalising on regularity of experiences – less straightforward than with cross-sectional data) Qualitative study as supplement for empirical enrichment (understanding motivations for transitions) or conceptual development (e.g. social capital)
Types of mixed method longitudinal: full or parallel approaches Quantitative data collection as part of ongoing ethnographic study Qualitative components to LS (e.g. collection of verbatim responses to more open questions, vignettes) Repeated observation on same individuals using survey and qualitative methods: presents considerable opportunities and challenges?