Possible Ways to Blend Qualitative and Quantitative Research Methods Ian McDowell Seminar Presentation 1997 Overview of Talk • Research methods gradually evolving in recognition of inadequacies in current methods • Two paradigms: positivist / quantitative vs. subjectivist / postmodern • There are strengths in each … • Types of blend of the two, in: – – – – study design data collection analysis comunicating results Styles of Thought (how do we know that we know what we think we know?) Many dualisms throughout history of thought: • Quantitative vs. qualitative • Deductive vs. inductive • Right brain and left • Yin and Yang • Apollonian and Dionysiac • Male and female • Reductionist vs. systems thinking Changing philosophies of knowledge • 17th & 18th centuries: order, logic and science, world seen through senses. Mechanical world. Realism and logical positivism • 19th century - social revolution: can we analyze behaviour logically? Idealism: the human mind as source of knowledge; people, not logic, crucial in explaining reality • 20th century - phenomenology; qualitative research Two paradigms • Challenge: biological variability – should we focus on the general or the specific? • General = public health; epidemiology; deduction. Nomothetic • Specific = clinical medicine; psychology; induction. Idiographic Quantitative approach • Describes and imposes external structure on data • Gives parsimonious summary of results: reductionist (for example, shared variance is attributed to one variable; hence it is reductionistic) • Seeks to isolate systems from their enviornment and to generalize • Efficient, but incomplete view of interconnectedness of reality • Asks the “How?” question • Externally valid Qualitative approach • Interprets, explains; generates concepts • Seeks to be open, flexible • The investigator is the instrument; art versus science • Sampling becomes a crucial issue (in data collection and in analysis) • “Somewhat magical approach to analysis” • Asks the “Why?” question • Particularizes; internally valid Blending Qualitative and Quantitative • Metaphor of binocular vision • Seeks to array strengths of one against limitations of the other • Nature of the balance may depend on stage of study: qualitative in a process evaluation, quantitative in outcome study, for example Five blends Hierarchical model: one method takes the lead i. ii. qualitative leads quantitative leads Partnership model: equal but contrasting contributions iii. sequential iv. cyclical v. simultaneous application (triangulation) Applying these Types of Blend In different stages of research: 1. Conceptualizing the study 2. Collecting data 3. Analysing data 4. Interpreting the data Stage 1: Conceptualizing the Study • Hierarchical model, quantitative leading, in “hard” studies • Hierarchical model, qualitative leading, in “soft” topics • Partnership model applicable in mixed studies or in broad programme of research that involves sequence of individual studies • Sequential partnership in formulating study: qualitative leads into quantitative Stage 2: Collecting the Data • Goal of blending approaches is to compensate for limitations in each approach • Hierarchical model illustrated by data supplementation (e.g., qualitative interviews highlight responses to a standardised questionnaire) • Partnership sequential model illustrated in qualitative work to develop questionnaires Stage 3: Data analysis • Generally hierarchical; determined by design of study. Orientation of funding agencies often makes it hard to achieve a true balance (“disciplinary racism”) • Hierarchical, quantitative leading illustrated by analyses of outliers • Hierarchical, qualitative leading: case studies are followed by secondary analysis fo quantitative data (e.g. a survey) to estimate representativeness of insights gained from the case study • Iterative analyses in partnership model, but this will probably be criticized from both camps. 4: Interpretation & Dissemination • Hierarchical, quantitative leading: – Use case histories or quotations to illustrate quantitative results – Use qualitative results to comment on exceptions to the rule • Hierarchical, qualitative leading: use quantitative results to validate what people suspected all along Future Directions • There’s increasing awareness in funding agencies of the importance of qualitative research. It’s a start, but…. • However, the paradigms are sufficiently different that it’s very hard to blend them: attempts rapidly lead to criticism that you are perverting the tenets of each approach • Disciplinary purity seems remarkably important to academics – a fundamental part of personal identity – so conflicts will be common • A successful blend will be truly “transdisciplinary”