Presentation on theme: "Multi-Methods in Longitudinal Research Bren Neale and Janet Holland (University of Leeds, London South Bank University)"— Presentation transcript:
Multi-Methods in Longitudinal Research Bren Neale and Janet Holland (University of Leeds, London South Bank University)
Research Practice Principles of research methodology cannot be allowed to stand on their own but must always be figured in relation to practice. …The world of research practice is multi-sited and multi- dimensional in its substance. … [hence] it is unwise to privilege any form or level of social reality over another. … We should be loath to seek ultimate knowledge of the world from [one source], whether a theoretical construct such as social structure or the seemingly authentic inner world of the individual [Seale et al 2004, 1-4]
Choosing methods The choice of research design and method is in practice likely to depend upon the nature of the research problem … how important is it to be able to describe a particular social process or to create a typology of different modes of relationship? How far is it the studys aim to test a hypothesis…[or] to find out how people think about a particular phenomenon?... The aims and questions posed by many research investigations are varied and cannot be addressed by a single research approach or strategy … a multi-method approach is therefore likely to be called for. [Brannen 2004]
Integrating or complementing? Different types of data need to be seen as constituted by the assumptions and methods that elicit them. In this perspective, qualitative and quantitative data need to be treated as broadly complementary, though not necessarily as compatible, rather than as adding up to some rounded reality, as advocated by exponents of triangulation (Brannen 2004: 313)
Longitudinal Research: theoretical drivers Researchers have developed new modes of enquiry that take account of the dynamic nature of peoples lives … Approaches to thinking dynamically have triggered the beginning of an intellectual revolution, one that blends insights from across the social sciences, merges quantitative and qualitative methodology and combines macro and micro views of society… Great questions of our time are disentangled by linking social change to the changing life course patterns of individuals and vice versa. … Modernity implies the institutionalisation of the individual the self and the life course as new social entities [Leisering and Walker 1998 The Dynamics of Modern Society xiv. 6-7; also Ruspini 1999).
Definitions and Designs Comparing temporal data, usually data generated over time. Time is built in in a variety of ways, e.g. data collection over short or long periods of time, or continuous data collection using ethnographic methods (Menard 1991, Ruspini 2002) Methods: Repeat cross sectional surveys/replication methods (different samples at follow up, for analysis of net changes for particular populations) Prospective panel methods: tracking same samples, e.g. a birth or generational cohort, family or other social groups over time using an array of methods to discern continuities and changes as individual or linked lives unfold. Retrospective methods: generating historical materials and records, life histories/life event data
Varied integrative Strategies combining structured and open data collection simultaneously or sequentially within a single study Combining insights across different and complementary studies Extensive (quantitative) framework –drawing on intensive in depth data/insights to contextualise and enrich the analysis and findings Intensive (qualitative) framework – drawing on extensive, large scale data/insights to contextualise and enrich the analysis and findings.
Laub and Sampson: Crime over the life course 1. Unravelling juvenile delinquency Sheldon and Glueck, Three waves of structured interviews and public records/statistical data collection (age 14,25, 32) to investigate the lives of 500 delinquent boys from poor neighbourhoods in Boston: Extensive data sets 2. Laub and Sampson secondary analysis, 1995 based on the combined data sets Laub and Sampson follow up study, including life history narratives with 52 delinquents, now aged 70, 2003.
Laub and Sampson (2003) Aims of the research: to unite qualitative and quantitative data to illuminate the processes of human development and continuity and change in criminal behaviour over the life course…. The integration of qualitative and quantitative data is notoriously difficult in practice and there are few guides to succesful examples (2003: 1)
Laub and Sampson Analysis Constructed detailed event histories from the structured and factual data and used a variable frame and statistical techniques to analyse the data. Findings: Marital and employment status influences criminal behaviour over time Used person frame and narrative analysis with a subset of 70 life history cases. Theoretically sampled using the quantitative data as a sampling frame: focus on high and low employment and marital status with varied criminal activity to assess quant findings. Discovered other salient factors, e.g. alcoholism or poor marital relationships counteracting employment and marital status, and military service as a stabilising factor. Criticised for failing to reveal the inner logic of these lives
Laub and Sampson follow up study Followed up the sample of 500 through accessing criminal and death records to discern what had happened to them, and updated the original data set. Based on the updated records they then tracked down and re-interviewed 52 of the surviving men, now aged 70. Theoretical sampling frame was based on the new data available to them; they sampled for criminal persisters, desisters, intermittents, and zig-zags. Used a life history calendar to collect event data (number, timing, sequence, and duration of marriages, divorces, employment, residence and so on), complemented by a focused retrospective life history interview
Laub and Sampson 2003 Defined strategy as a systematic weaving back and forth between numerical and life history data, in order to enrich the overall analysis e.g. life histories revealed the importance of turning points in explaining nature of changes in criminal activity in the mens lives, e.g. the significance of getting married, entering the military. They were then able to further interrogate these themes in the quantitative data They used both life history and statistical presentation of their findings. This enabled a more detailed understanding of individual lives. The narratives revealed marked variations in the life trajectories of the men and multiple pathways to the same outcome.
ESRC Young Lives and Times: The dynamics of young peoples relationships Qualitatively driven longitudinal study October Age cohort of 40 young people, tracked over 10 years, using rich array of ethnographic methods, framed by regular waves of in depth interviews. Aim to create a cultural inventory of young lives Links to innovative local survey (n=400) exploring social processes Links to DfES longitudinal study of young people (20,000 tracked over a decade); and BHPS youth survey data Iterative strategy, over time, for using complementary micro, meso and macro level data. Methodological as well as substantive study, with methodological refinement as the study progresses.
Challenges Although the distinctions often asserted between qualitative and quantitative research emerge as less robust under close examination, none the less clear distinctions exist between qualitative and quantitative researchers. Indeed, the distinctions look set to widen (Brannen 2004: 313)