Bolashak - Week 6 - CAQDAS with Dedoose

Slide 1

CAQDAS with Dedoose

(Computer-Aided Qualitative Data Analysis System)


Slide 2

Case study: How do EFL university teachers deal with the heterogeneous classroom?

“The present case study aims to investigate the strategies that Kazakhstani university English language teachers use to cope with heterogeneous students in their classroom.
To this end, two university English teachers in one of the higher institutions in Kazakhstan were observed during their teaching sessions and interviewed in order to identify the reasons why they used particular teaching strategies to deal with heterogeneous students. “
We are dealing with two interview transcripts.
Can also imagine some initial codes: strategies; heterogenous students; reasons. Maybe also: classroom.


Slide 3

QDA @ UIUC

Lots of links & resources @ the Library:
https://guides.library.illinois.edu/qualitative/home


Slide 4

Keep a separate secure Excel file, with codes and names Rename files with codes, using a consistent file name convention Rename participant in the transcript: “Aizere” > “Participant A” Use Microsoft Word’s Protect Document feature to remove names from properties Take care with identifying details FIRST, ENSURE CONFIDENTIALITY

Image



Image

Image


Slide 5

Keep a separate secure Excel file, with codes and names Rename files with codes, using a consistent file name convention Rename participant in the transcript: “Aizere” > “Participant A” Use Microsoft Word’s Protect Document feature to remove names from properties Take care with identifying details FIRST, ENSURE CONFIDENTIALITY

Image



Image



Image

Image


Slide 6

Braun & Clarke (2006) - from week 2

familiarising yourself with your data
Search for ‘latent or semantic themes’
‘Take notes’ or mark ‘ideas for coding’
‘Transcribing verbal data’
generating initial codes
‘writing notes’, highlighting text (different colours = different themes / codes)
searching for themes
Analyse codes
Tables, mind-maps, ‘theme piles’
reviewing themes
Develop & refine candidate themes
Data within one theme should be ‘homogenous’; data from different themes should be ‘heterogenous’
Refine (i) by document and (ii) across the whole corpus
defining and naming themes
producing the report


Slide 7

Steps 1, 2 – Sample THEMES & codes

Image

Some key terms?
“Difficult”
“Lesson plan”
“Bored”, “Distracted”, “Know everything”
“Challenge”



Draft** Themes and Codes
Lesson plan
Task Difficulty
Common
Difficult / Challenging
Student State**
Arrogance
Boredom
Distraction


Slide 8

Try codes with chatgpt, deepseek

Can codes be extracted? We can test this out…


Slide 9

Choosing a qda product

Commercial vs Open Source?
Manual vs Automated Coding?
Dedicated Tool vs General-purpose Product?
QDA (Qualitative Data Analysis) is a niche product – not used outside of academia (though has applications in marketing – see DoveTail)
See Week 3 – encourage use of AI, but with human checking

Also: remember why we want to do coding. Many studies assume the following is acceptable as a methodology:
“We apply thematic analysis (Braun & Clarke 2006) to transcripts of interviews / focus groups”
But why? When will other approaches work:
Read & highlight the text for meaningful responses to questions
Word frequency analysis / word clouds / other NLPs techniques (using Python & ChatGPT - week 3)


Slide 10

Thematic analysis – pause

Sometimes “coding” is a fancy way to lose meaning in your data. Make sure you know why thematic analysis is right
If data is small (1-2 interviews) – maybe simply read & comment?
If data is large (100s+ of documents) – will thematic analysis scale? Need algorithmic support, alternate techniques
Common arguments about “how many” interviews / surveys – determined by technology & research time – not necessarily epistemological concerns (validity, reliability)


Slide 11

Image Nvivo


Slide 12

Atlas.TI

Image


Slide 13

MaxQDA

Image


Slide 14

Dovetail

Image


Slide 15

Image Dedoose

+1 month free trial
Collaborative – very useful when working in teams


Slide 16

Image Open source / free?

Not really?
Products like QualCoder, Taguette - no experience with them, but seem individual hobby projects
Would encourage trying them out
Require more technical expertise? Less fully featured?
Not like quant data analysis: R, Python both hugely popular, supported, etc
AI: definitely an option (see Week 3). Many QDA products include an “AI” option at additional cost (likely a “wrapper” around ChatGPT / Claude).

Image


Slide 17

Microsoft word?

Fine for small projects
Requires discipline in colour coding / highlighting
Hard to extract metrics: common themes / codes
Not good for collaboration: hard to check inter-rater reliability (do Raigul and Aigerim code data in the same way?)

Image


Slide 18

dedoose OVERVIEW

Benefits: Cheap, simple, collaborative
Data is based in the cloud – not as private as desktop-based
NOT an endorsement.
All QDA software is quite complex, buggy… Not an area of “high” quality software – 1990/2000s era with incremental updates.
Expect to spend time re-doing some steps, backing up data, learning tutorials…

Very quick overview – for more information:
Download:
https://www.dedoose.com/resources/articledetail/dedoose-desktop-app
Lots of videos, tutorials etc:
https://helpdesk.dedoose.com/hc/en-us
https://www.dedoose.com/home/resources

Explore the Demo Project…


Slide 19

Image


Slide 20

Image Steps

Sign up online
Download and open Dedoose software
Login
Go to Projects tab
Click “Create Project”
“Import Data”



Image


Slide 21

Image Image

7. Add codes: strategies; heterogenous students; reasons; classroom
8. You have two key concepts:
Codes
Media (documents)
9. Your job is then to open media, apply codes, add descriptors / fields and then analyze



Image


Slide 22

What if you need to classify? Image

Interviews by age, gender, other variables
Dedoose has concept of descriptors / fields / sets
You might want to test:
Does the code “strategies” appear more often among women than men participants?


Slide 23

In this example I have said Participant A is a woman, Participant B is a man (both are I think women in reality?) And there is a difference: participant B focuses more on difficulties, A on strategies So we could see how we begin to develop a picture of the data…

Image


Slide 24

Next steps?

More codes, sub-codes: Different kinds of strategies, difficulties
More fields (primary language – Kazakh, Russian?; age; urban / rural; school level, etc)
Add memos – reasons or comments about assignment of codes to media (excerpts) – can help with analysis
Explore different analytic combinations
Collaboration – examine inter-rater reliability
Mixed methods – multimedia, quantitative data etc


Slide 25

Next steps?

Convert into textual commentary in a journal article. Examples:
High level summary: “After coding the interview transcripts, we found teachers used a variety of strategies – pairing students, being more permissive of weaker students making errors, offering token rewards (e.g. chocolate bars), challenging high performing students, offering praise – and even singing songs!”
Include excerpts as quotes: “One participant (B) stated that pairing was helpful: ‘Then, somehow, these strong students they help weaker students.’”
Add** detailed analysis: “Teachers with varied strategies – such as participant B – rarely spoke of difficulties, suggesting that having a repertoire of such strategies was important in managing classrooms with heterogenous students”.
Examine
differences, possible causes, and potential solutions**: “[DIFFERENCE] Curiously, female teachers seemed to have more strategies for managing heterogenous classrooms than male. [CAUSE] This may be due to male teachers being focussed on elevating higher performing students, and suggests [SOLUTION] more effort needs to be spent on training teachers to focus on overall classroom performance.”