Research methods in clinical psychology: An introduction for students and practitioners Chris Barker, Nancy Pistrang, and Robert Elliott CHAPTER 12 Analysis, interpretation and dissemination
Overview Interpretation –What is the strength and significance of the findings? –What are their scientific and professional implications? Dissemination –Making the findings known
Qualitative analysis: overview Analysis is an inductive process Many different approaches –vary in depth of interpretation / inference –method should fit research questions and the data Generic processes / general principles Within-case and cross-case analysis
Frequently used approaches Grounded theory Interpretative phenomenological analysis (IPA) Discourse analysis Content analysis
Preliminaries to data analysis Transcriptions –different conventions –anonymity “Immersion” in the data
Generic processes in analysis identifying meaning categorising integrating Note: cyclical, not linear
Identifying meaning identifying and labelling ideas line-by-line (microanalysis) meaning units codes (labels): ‘in vivo’ v. abstract implicit v. explicit meaning
Categorising themes or categories method of ‘constant comparison’ “saturation”
Integrating linking themes / categories conceptual framework or hierarchical structure
Computer packages for qualitative analysis Good for sorting and searching, linking categories e.g., ATLAS-ti, NUD*IST
Writing up the results Different models –conventions for different genres of qualitative research –what best captures the essence of the data? –be guided by the research questions Narrative account –tell a story –describe the phenomenon –illustrate with examples Table of themes/ tree diagrams
Good practice in qualitative analysis guidelines for evaluating qualitative research, e.g.: –credibility checks –have the research questions been answered? –is the analysis coherent and integrated? Elliott et al. (1999); Willig (2001); Yardley (2000)
Quantitative approaches Measures of strength and significance of the findings
Statistical conclusion validity Was the study sensitive enough? –Large enough sample? –Error minimised in measurement and design? Do the variables covary? –Were the statistical methods appropriate? If so, how strongly? –Significance (Shadish, Cook & Campbell, 2002)
Significance of the findings Statistical significance Effect sizes Clinical significance
Statistical significance p-value (alpha level) of statistic –e.g., 2 (1) = 4.7, p = 0.03 null hypothesis testing framework currently controversial –replace with confidence intervals? value dependent on sample size
Effect size measure of magnitude independent of sample size depends on statistical test often classified into small, medium and large (see Cohen)
Effect sizes: Meta-analysis Pioneered by Smith & Glass (1977) Aggregates several studies, using effect sizes Advantages: –Quantitative effect size index –Can also examine study variables (e.g., investigator allegiance) However: GIGO (garbage in, garbage out)!
Clinical significance Measure of meaningfulness –do patients actually improve? –“endstate functioning” Jacobson and Truax (1991) –reliable change –clinical significance cut-offs “Number needed to treat” –used in evidence-based medicine
External validity Can the findings be generalised across: –persons –settings –times? Replication –Literal –Operational –Constructive
How research is used and interpreted Dissemination –research as a public activity –feedback to staff and managers –feedback to participants Publication Research utilisation –does research affect policy? –models of research utilisation (Weiss, 1986) Political issues