Qualitative Data Analysis 1

EDUC 894
Week 8
Plan for Today
Conference Debrief
 What’s Hot in Analysis Methods: Engaging Complexity
Group-work & Team Consultations
 Midterm Course Feedback Sheets
--------------------Dinner Break------------------ Group-work & Team Consultations cont.
 Qualitative Data Analysis (7 pm)
 Overview of Approaches
 Class Activity: Analyzing Data
Moving Forward
 Deliverable 3 Handout & Due Date Discussion
 Reading: Auerbach only (online book, pick up handout)
Qualitative Data Can Be:
 Observation notes
 Images
 created by researchers or subjects
 Student work
But one thing is always true:
Qualitative is voluminous
(that’s a nice way of saying that it is easy to get overwhelmed)
The Overriding Goal(s)
Make inferences from the data to
address your research questions
 Connect the empirical (specific data)
with the theoretical (research inquiry)
 Most importantly it is a process of
making meaning of the data
 This often involves looking for patterns
 This can occur simultaneously or in iteration
with data collection
 And what does the data mean to whom…?
Qualitative Data Analysis (QDA)
Approaches: 2 Schools
Back to the miner and the traveler
What are they looking for & how will they go
about it?
 Think about how their interview styles have
set them up for the data analysis stage…
QDA: The Miner Approach
“Testing Theory” (Glaser)
 Does the data support our
 Does the data fit a preestablished framework?
Closed Coding / Rating
 Interpretive
 Automatic
Image Source: http://www.synthstuff.com/mt/archives/bertrand-miner.jpg
QDA: The Miner Approach II
Qualities of a Good Analysis
 Objective & Reliable
○ Multiple raters
 Inter-rater Reliability
 Number of decision points
 Tolerance & Reconciliation
○ Multiple measures
 Rubrics and consistency
 Valid
○ How do the indicators used
serve as evidence of the ideas
you are looking for?
 Examines all data in the sample
Image Source: http://www.synthstuff.com/mt/archives/bertrand-miner.jpg
QDA: The Miner Approach III
Is this still really “qualitative”
data analysis?
Image Source: http://www.synthstuff.com/mt/archives/bertrand-miner.jpg
QDA:The Traveler Approach
“Generating Theory” (Glaser)
 What is the data trying to tell us?
 What are important themes related
to the topic / phenomenon / RQs?
 What structure emerges from the
Coding processes
 Range in structure and imposition of
 Creswell presents a generic
Image Source: http://www.hmoon.com/garments/cloaks/traveler1-detail.jpg
QDA:The Traveler Approach II
Coding process decisions
 By question or overall
○ This highly relates to the kind of
interview you conducted
 Target categories or completely
 Number of coders; how and when to
“Constant Comparison”
 Item to item
 Item to Category
 Category to Category
Image Source: http://www.hmoon.com/garments/cloaks/traveler1-detail.jpg
QDA:The Traveler Approach III
So when are you done?
 Unlike the Miner – you aren't’
focused on “leaving no stone
unturned,” but you will likely go over
all the data several times.
 You will constantly be revising,
combining and rethinking your
codes - when no new themes are
emerging, you may be nearing
“theoretical saturation”
 Or you may need to go collect more
Image Source: http://www.hmoon.com/garments/cloaks/traveler1-detail.jpg
Theorectical Saturation
Data Collection
Data Analysis
Close to Saturated
More Refined
Saturation of
QDA: The Traveler Approach IV
Qualities of a Good Analysis
 Confirmability (Results are
○ Triangulation
○ Search for Negative Evidence
 Dependability (Results are Stable)
○ Overlapping Methods
○ Cycles of Analysis and Data
 Credibility (Results are Plausible)
○ Detailed analysis process
○ Coherence of results
○ Discussion of alternative explanations
○ Member Checks
Image Source: http://www.hmoon.com/garments/cloaks/traveler1-detail.jpg
QDA: Some Final Thoughts
There is great variation in how QDA is done
Your approach should be specific to your study
(your RQ, your data and how it was collected)
 Since each instance is unique, it is very important to
document your process
The two approaches described here (generating
theory and testing theory) can be combined
(Glaser’s top left quadrant)
 You’ll read about another way of doing this in
 Next week we’ll look at one technology to help
manage masses of qualitative data (Atlas-ti)
 Class next week will be in
Room 3100
Analyzing Data Activity
 RQ: Are X,Y & Z
important issues
about vacations for
graduate students?
 RQ: How do
graduate students
think about
Data source = X letters written by grad students about their vacations.
Working in pairs, pick a stance and RQ and think about how
you will analyze the data. In particular make sure to consider
how the two of you will coordinate your efforts and how you
will be able to justify either the validity & reliability (objectivist)
or trustworthiness (constructionist) of your analysis.