Forman Qualitative Research for Health Services Re

Research Design and Analysis 710; Fall 2005 Qualitative Research for Health Services ResearchersSession 2Designing Qualitative Research: 

Research Design and Analysis 710; Fall 2005 Qualitative Research for Health Services Researchers Session 2 Designing Qualitative Research Jane Forman, ScD, MHS Ann Arbor VA HSR&D

Objectives: 

Objectives Framing research questions Study design: matching methods to research questions. Understand how to provide a rationale for sampling choices. Understand various sampling strategies.

Qualitative Research Questions: 

Qualitative Research Questions Research question: a question that the research is designed to address, vs. questions an interviewer asks an interviewee. Focus: purpose, available resources, time, expertise Don’t think in terms of variables that you’re trying to relate to each other. Do think in terms of identifying factors and understanding how they work. Questions can evolve.

Research Strategy and Design: 

Research Strategy and Design Research design is not an entire advance blueprint; we make decisions throughout the research process But, a detailed design from which to start is very important and useful. You need a rigorous plan. Think widely and creatively, then pare down given best fit with the essence of your inquiry, and constraints on it Talk through methods, techniques, and data source issues with others

Research purposes: 

Research purposes Basic research Applied research Summative evaluation Formative evaluation Action research

Matching Methods to Research Questions: 

Matching Methods to Research Questions How and why particular methods, techniques, and data sources might yield data which will help you to answer your questions (vs. I know how to do interviews so I’ll do them) (Mason p.27) What data sources and methods of data generation are potentially available or appropriate? What can these methods and sources feasibly tell me? How or on what basis do I think they could do this? Which of my research questions could they help me to address? Which elements of the background (literature, theory, research) do they relate to?

Matching Methods to Research Questions What constitutes knowledge or evidence relevant to my research questions?How do I go about generating such knowledge and evidence?: 

Matching Methods to Research Questions What constitutes knowledge or evidence relevant to my research questions? How do I go about generating such knowledge and evidence? (Mason 2002, Ch.2)

Units of Analysis: 

Units of Analysis Examples: Individual people, program, community, organization “The key issue in selecting and making decisions about the appropriate unit of analysis is to decide what it is you want to be able to say something about at the end of the study.” Patton, p.229

Sampling: Representational vs. Purposeful: 

Sampling: Representational vs. Purposeful Representational Goal: Enable generalizations from study samples to populations. The sample displays variables (e.g., age, gender) in similar proportions and patterns to the total population about which you wish to make generalizations. Statistical conventions are used to calculate the probability that patterns observed in the sample will exist in the wider population.

Sampling: Representational vs. Purposeful: 

Sampling: Representational vs. Purposeful Purposeful Goal: To understand a phenomenon, not to represent a population. The selection of information-rich cases for intensive study. Commonly used in qualitative research.

Selected Types of Purposeful Sampling: 

Selected Types of Purposeful Sampling Criterion Maximum variation Stratified Typical case Deviant case Random Theoretical

Criterion Sampling: 

Criterion Sampling A generic form of sampling involving the selection of cases on pre-conceived empirical or theoretical criteria. Sometimes called selective sampling. Criteria may entail other kinds of sampling, such as typical or deviant case sampling. Criteria can be a score on a measure (qual/quant combined study).

Maximum Variation: 

Maximum Variation Selection of cases to represent a set of variations on specified aspects of a phenomenon (variables) that are relevant to your phenomenon of interest. Equality of number is not relevant. What is important is that you have information so you can say something about that variable. Phenomenal, not demographic, variation Phenomenal variation can be demographic Researchers choose which and how many features to vary (heterogeneity) and which features to fix (homogeneity). More variation = larger sample size. Most appropriate for qualitative descriptive studies, or beginning GT studies.

Stratified Purposeful Sampling: 

Stratified Purposeful Sampling Selection of cases showing combinations of pre-selected variables. Can use stratified purposeful sampling in grounded theory when there’s already theory development.

Random Purposeful Sampling: 

Random Purposeful Sampling Used when there is a very large pool of potentially information-rich cases and no obvious reason to choose one case over another.

Typical Case Sampling: 

Typical Case Sampling Selection of cases conforming to a specified standard, or pre-conceived norm. Comparatively typical: Within the sample you have selected, they are typical. Normatively typical: Typical in relation to something out there, e.g., a measure.

Deviant Case Sampling: 

Deviant Case Sampling Selection of cases departing from a specified standard, or pre-conceived norm. Comparatively atypical: Within your study, something stood out. Normatively atypical: Not typical according to a standard outside your study.

Theoretical Sampling: 

Theoretical Sampling Selection of cases to complete a theoretical line (I.e., to achieve theoretical saturation) developed in the course of the study. Theoretical vs. empirical case.

Sample Size: 

Sample Size Sample should be large enough to make meaningful comparisons in relation to your research questions. Factors that affect sample size: Number of comparison groups (more comparisons  larger sample size). Detail, complexity and depth (more detailed, complex, and in-depth  smaller sample size) Develop explanations to account for similarities and differences in particular contexts. Seeing nothing new in newly sampled units involves recognizing what is there and what can be made out of the data already collected, and then deciding whether it is sufficient to create an intended product. There’s a tendency to over-sample in the health sciences. Underestimation of how much you get out of the data.

Sampling Decisions: 

Sampling Decisions Sampling is part of the detailed design that you develop before you begin your study. You make sampling decisions during data collection based on what you find in your data.

Keep Asking These Questions: 

Keep Asking These Questions Why is this or that category or group relevant? In what ways would including it or them in my study help me in developing the overall kind of explanation I wish to develop, or in understanding the process I wish to understand?

References: 

References Michael Q. Patton, Qualitative Research & Evaluation Methods, 3rd ed., Sage, 2002 Jennifer Mason, Qualitative Researching, Sage, 2002 Margarete Sandelowski, series of articles on qualitative research in Research in Nursing and Health. Marshall and Rossman, Designing Qualitative Research, 3rd ed., Sage, 2002