10: Mixed Methods Flashcards
List five reasons why mixed methods are useful.
Triangulation: find convergence & agreement across different methods; avoids method variance, increases convergent/divergent validity.
Complementarity: develops enriched, elaborated explanation of a given phenomenon using strengths of both methods.
Development: first method used to help develop second method.
Initiation: find inconsistencies, contradictions, paradox between methods.
Extension: expand breadth or range of inquiry by using different methods for different inquiry components.
Post-positivism accepts what and rejects what? How does it view science?
Accepts notion of objective reality outside of our senses that we can study.
Rejects idea of incommensurability - assumes there is one true way of seeing the world.
All types of measurement and theories fallible; goal of science to find “truth” about reality while acknowledging we never will.
Social constructionism accepts what and rejects what? How does it view science?
Accepts idea of incommensurability - assumes there are multiple, valid ways to understand world.
Rejects notion of objective reality outside senses that can be studied.
Science should focus on how arguments are constructed and to what end.
Describe the advocacy-transformative viewpoint. What does it focus on?
Doesn’t care about philosophical concerns like objective reality or incommensurability; research politically involved.
Power differentials, tries to empower and give voice to stigmatized groups.
Describe the pragmatism viewpoint. How does it view science?
Doesn’t care about philosophical concerns like objective reality or incommensurability.
Purpose of science is to be useful, or practical in day-to-day life.
Per pragmatists, the only thing that matters is the practical, real-world utility of research findings. What is the term for this?
Instrumental truth.
What are the three steps in designing a mixed methods study?
- Choose a theoretical lens.
- Implement and prioritize data collection (i.e., concurrent, sequential implementation; equal, unequal priority).
- Data integration.
Describe the prioritization of data collection in a mixed methods study using concurrent implementation.
Concurrent implementation -> equal priority (QUAN + QUAL) / unequal priority -> QUAN priority (QUAN + qual) / QUAL priority (QUAL + quan).
There are three possibilities as to when data integration will occur. What are they?
Analyze quantitative/qualitative data separately, compare & contrast findings.
Convert qualitative data into quantitative codes, analyze both together in quantitative analysis.
Connect two types of analyses together in separate studies (e.g., pilot qualitative study which informs subsequent quantitative).
In sequential explanatory designs, quantitative data is collected first. What are the three possibilities following?
Equal priority to both data types: QUAN → QUAL.
Priority given to qualitative data: quan → QUAL.
Priority given to quantitative data: QUAN → qual.
Describe sequential explanatory design. How does it use qualitative and quantitative data?
Data analysis is connected; typically, qual portion is a selected subsample of the QUAN data.
Qualitative data is used to explain the quantitative results.
Describe sequential exploratory design. Why is it used?
Data analysis is connected (QUAL collected first); typically, the quan portion may or may not involve the same people as the QUAL portion.
Used to explore poorly measured or conceptualized construct, creating new test, or to increase external validity of QUAL results.
Describe the two phases used by Mihretu et al. (2017) in their sequential exploratory design.
Phase 1: sample of 11 khat users participated in in-depth interview. 4 focus groups with n = 9 per group collected. Data analyzed from phenomenological approach.
Phase 2: checklist of addiction indicators developed from interviews, administered as closed-ended questionnaire to n= 102 Khat users.
Describe concurrent design.
Both types of data collected at the same time (QUAN + QUAL). Analyze both sets of data, compare & contrast. Ideally, both types of analysis produce similar results.
When is a concurrent design “nested”?
Place more emphasis on one type of data.
