Call Girls In Doddaballapur Road ☎ 7737669865 🥵 Book Your One night Stand
Mixed methods in social and behavioral sciences
1. Mixed methods in social
and behavioral sciences
ARNOND SAKWORAWICH, PH.D.
2. Origin of mixed methods
MTMM (Multi-trait multi-method)
Campbell and Fiske (1959)
-Convergent validity (Same trait, different methods →relatively high
correlation)
-Discriminant validity (Different traits, same method → low correlation)
-Method variance (Correlation among different traits and same method
is way higher than correlation among different traits and different
methods)
3. Origin of mixed methods
Triangulation
From military (and history)
-Time, space, context, motivation
Instrument Development
- Critical incident technique (Flanagan)
Behavioral Event Interview (Competency modeling by McCleland)
4. Positivism versus Phenomenology
• Quantitative
• Empirical data
• Scientific approach
• Try to measure and
quantify
• Study the issue by
isolating (et cetera
puberis)
• Aims at uncovering
causal relationships
among variables
• Qualitative
• Experience, insight,
intuition
• Meaning and
interpretation
• Connectedness and
complexity
• Holistic approach
• Aims at describing and
understanding
phenomenon
5. Why we need mixed methods?
- Dialectical perspective
◦ -Confirmation, disconfirmation
◦ - Wearing different eyeglasses, Can we still see the same thing?
◦ - What make the differences and similarities?
- Pragmatic perspective
◦ - Whatever works best for results?
◦ - Use different tools suitable for different job
◦ - Strengthen results by compensating weaknesses from one method
7. Fundamental components of mixed method design
1. Timing
◦ -Concurrent
◦ - Sequential
◦ Quan then Qual
◦ Qual then Quan
2. Weighting
◦ Equal weight
◦ QUAN>qual
◦ QUAL>quan
3. Mixing
◦ -Connecting
◦ - Integrating
◦ -Embedding
4. Theorizing
◦ Explicit
◦ Implicit
8. Fundamental components of mixed method design
3. Mixing (between QUAN and QUAL results)
- Integrating
◦ Quantify qualitative data into quantitative data
◦ Counting, Content analysis, Frequency distribution
◦ Coding then statistical analysis
◦ Text mining
◦ Inference and disputed authorship
◦ – Federalist paper (85 papers)
◦ Alexander Hamilton
◦ James Madison
◦ John Jay
◦ Promoting ratification of the USA
9. Fundamental components of mixed method design
3. Mixing (between QUAN and QUAL results)
- Integrating
◦ Qualify quantitative data into qualitative data
◦ Exploratory factor analysis to investigate how variables hang together!
◦ Naming factor – make it interpretable
◦ First factor, modernism, consists of haughty, modern, lavish, clean, disciplinary, self-
confident, creative, courteous, romantic, clever, competent, and independent. Five
countries with highest score on this factor are England, France, Japan, Italy, and USA,
while five countries with lowest score are Iraq, Mexico, Brazil, Iran, and Chile.
◦ Second factor, need for power encompasses tricky, invasive, authoritarian, exploitative,
selfish, brutal, aggressive-minded, nationalistic, and brave. Five countries with higher
score on this factor are USA, Iraq, Myanmar, Iran, and Israel, while five countries with
lowest score are Ethiopia, Egypt, Italy, Singapore, and Chile.
10. Fundamental components of mixed method design
3. Mixing (between QUAN and QUAL results)
- Integrating
◦ Qualify quantitative data into qualitative data
◦ Third factor, namely work value, comprises avaricious, diligent, patient, economical, boisterous,
and business-minded. Five countries with highest score on this factor are China, Japan, Vietnam,
Taiwan, and Singapore, while five countries with lowest score are New Zealand, Australia, France,
Venezuela, and Italy.
◦ Fourth factor, namely positive relationship, is formed of generous, kindhearted, gregarious,
pacific, simple, and honest. Five countries with highest score on this factor are New Zealand,
Australia, Laos, Vietnam, and Canada, while five countries, while five countries with lowest score
are India, Ethiopia, Saudi Arabia, Russia, and France.
◦ Fifth factor, namely less developed, is composed of stupid, lazy, poor, and dirty. Five countries
with highest score on this factor are Laos, Cambodia, Myanmar, Ethiopia, and India, while five
countries with lowest score are Brazil, Australia, Mexico, Israel, and Taiwan.
◦ Sixth factor, namely happy go lucky-sober, includes enjoyable and pious.
◦ Five countries with highest score on this factor are USA, Brazil, while five countries with lowest
score are Iran, Iraq, Israel, England, and Vietnam.
11. Fundamental components of mixed method design
3. Mixing (between QUAN and QUAL results)
- Integrating
◦ Qualify quantitative data into qualitative data
◦ Cluster analysis to investigate how people or object are clustered together
12. -First cluster, non-salient countries
-Latin American sub-cluster
- Venezuela, Argentina, Brazil, Mexico, and Chile
-Developing sub-cluster
-Malaysia, Russia, South Korea, Germany, and Taiwan
-Second cluster, underdeveloped countries
-India, Cambodia, Myanmar, South Africa, Saudi
Arabia, and Ethiopia.
-Third cluster, warrior
-Iraq, Iran, and Israel.
-Fourth cluster, developed countries
-Italy, Singapore, Japan, France, and England
-Fifth cluster peaceful countries
-Australia, New Zealand, Canada, Laos, and Vietnam
13. Fundamental components of mixed method design
3. Mixing (between QUAN and QUAL results)
- Connecting
◦ - Results from QUAN are used in QUAL
◦ - Results from QUAL are used in QUAN
◦ Critical incident technique
◦ Performance appraisal form construction
- Embedding
◦ QUAN supports QUAL
◦ QUAL supports QUAN
14. Fundamental components of mixed method design
4. Theorizing
- Explicit
◦ Used theory (ies) or lens to guide data collection and analysis
◦ Critical theory, Postmodern, Feminist, PAR, Empowerment etc.
- Implicit
◦ Bracketing oneself!
16. 1. Sequential Explanatory Design
- Popular among quantitative researchers
- Explain quantitative results especially unexpected results
- Typically weights QUAN more than qual.
- Mixing through connecting
- Explicit theory may or may not inform the overall procedure.
17. 1. Sequential Explanatory Design (Example)
Hawthrone Experiment
-Illumination and relay assemble
-The darker light, the higher productivity
- Confounding
- Follow-up interview
- Human relation movement
- Elton Mayo
- Human are not machine, we need love
and care!
18. Another examples
-Psychological tests and assessment were administered.
- Extreme scorers were followed up by in-depth interview and
longitudinal study.
- Study of psychopathology and its prognosis and treatment.
19. Another examples
- Industrial survey
- Benchmarking
- Case study (success and failure, best practice and worst practice)
20. 1. Sequential Explanatory Design
- Easy to implement
- Easy to describe and report
- Ex post facto (if QUAN is experiment)
- Lengthy duration
21. 2. Sequential exploratory strategy
- Appropriate to study something new and we don’t know much about it.
- QUAL →quan
- Typically weights QUAL more than quan.
- Usually data are mixed through connecting. QUAL data are mixed into quan
data in a later phase.
- Design may not be implemented through theoretical perspective.
22.
23. 2. Sequential exploratory strategy
- Item development
- construct clarification through qualitative research
- Item are from either critical incidents or interview transcription
- Survey
- Item analysis
- Reliability analysis
- Validation
- Big Five and Lexicon approach
24. 2. Sequential exploratory strategy
- Easy to implement
- Easy to describe and report
- Appropriate for construct / theory clarification
- Appropriate for model/theory building then model/theory testing
- QUAN can help increase external validity and generalization to another
population
- Take times
25. 3. Sequential transformative strategy
- QUAL → quan or QUAN→qual
- Mixing through connecting
- Theoretical framework is used to guide overall research design.
26.
27.
28. 4. Concurrent triangulation strategy
- Compare different approach to investigate convergence/dissimilarities
- Confirmation, disconfirmation, cross-validation, corroboration
- Typically weights equally
- Usually mix data through integration/Side-by-side comparison
29.
30.
31. (1) The group counselling offered us support which was different from the support we received from friends and family.
(4) It takes time to adjust to being HIV + and group counselling provided us with the time and opportunity to come to terms
with our new status.
(5) Group counselling was the most effective method we found in changing our behaviour. We needed the support and
companionship of the group to help us change our behavior to live effectively with HIV + .
(8) The group counselling has allowed us to make important changes in our lives. These changes have been difficult and without
the support of the group members we would not have been able to make them.
(9) Meeting with other HIV + people has provided us with a level of understanding that we have not found anywhere else.
(15) Admitting publicly that we are HIV + is a big step and concerned us all. Counselling can help the individual at this point and
provide support to reduce the level of stress. We recommend that individuals should be free to decide when the appropriate
time to go public is. However we recognize that inevitably others will come to suspect and we recommend that these people
help as much as possible when that time comes.
(17) The stress and anxiety that 'miracle cures' cause is counter productive to our health. We would rather find ways of living
positively with AIDS than relying upon drugs. We do not want to act as guinea pigs in drug trials and think that it is wrong of
researchers to excite the possibility of a cure when this is not a responsible or realistic possibility.
(18) People with AIDS is a better term than AIDS victim, but we would prefer just to be considered as people. We recognize that
everybody has to die at some point and we are no different. We want to be allowed to live our lives as completely as possible.
(19) We accept that AIDS is a terminal illness, but we suffer more shock, bitterness and fear because society stigmatizes and
discriminates against us.
35. Rationale
-People act as they believe.
- Picture inside his head = stereotype
- We use stereotype as cognitive framework and heuristics to decide and
generalize from one object to other objects.
- Stereotype is not accurate. It can cause prejudice and bias as well as conflict.
- Hostile stereotype → hostile foreign policy!
- Stereotype studies during cold war in Thailand
- 66/2523 policy agenda – Politics lead forward the military.
- About 10 Years of Taksin regime
- Red shirt (UDD), Yellow shirt (PAD), and PDRC
- The advent of coup d’etat and NCPO
36. Research objectives
-To study the characteristics of stereotypes towards various Thai political
groups as well as its leaders i.e. UDD, PAD, PDRC, NCPO, Taksin Shinawatra,
Yingluk Shinawatra, Sutep Theugsuban, Abhisit Vejajiva, Sondhi Limtongkul,
Prayuth Jantraocha, Democrat Party, Puerthai Party, and Thai monarchy.
- To study factors affecting those stereotypes, specifically media exposure, past
political behavior, subjective norm, and their beliefs.
- To design an intervention to mitigate the extreme stereotypes among Thais in
order to harmonize Thai society.
38. Three phases of research project
-Phase I: Qualitative research
- Key informant in-depth interview
- PDRC, UDD, PAD, Democrat Party, Peurthai Party- leaders
- People who participate in demonstration (UDD, PDRC, PAD)
- Converter
- Theoretical sampling (and snowball sampling)→Theoretical saturation
-Data analytic strategies – grounded theory and phenomenology
- Lists of adjective describing stereotype towards political groups
- Grounded theory of stereotype towards political groups
39. Three phases of research project
-Phase II: Quantitative research
- large scale survey and correlational research
- Multi-stage sampling
- Instrument development
- Data analytic strategies
◦ -Factor analysis – grouping attributes
◦ -Cluster analysis – grouping people
◦ - Multidimensional scaling - perceptual mapping
◦ - Log-linear model – Relate cluster to demographic variables
◦ - Profile analysis via multidimensional scaling – Stereotypical profiles
◦ - Dominance analysis – ranking beliefs and subjective norms predicting stereotype
◦ - SEM - modeling sterotype
40. Three phases of research project
-Phase III: Experimental research
- Design intervention to change the extreme stereotypes
- PAR
- Quasi-experiment
O1 X O2
O3 O4
-Repeated measure design
-ANCOVA