MAXQDA Free Training for Beginners

Whenever we want to conduct thematic analysis using MAXQDA, there is one thing that we should always do: have a plan.

We should always have a framework that we intend to use during the analysis.

One of the most popular frameworks that was popularized by two scholars is the Braun and Clarke six_steps framework of thematic analysis which includes:

• Familiarizing with the data
• Generating initial codes.
• Developing themes.
• Reviewing themes.
• Defining and naming themes.
• Writing the report.

This is one of the most common approaches I use to conduct analysis.

In this case, MAXQDA is just a tool that will help me manage my data; it is not the main reason behind my analysis.

 I organized my analysis based on a clear framework, such as the Braun and Clarke six-step framework of thematic analysis.

 Having known this, we can now open MAXQDA.

 As usual when we open MAXQDA, we will see some menus.

There is atop menu that has the main tools like:

• Home
• Import
• Codes
• Menus
• Analysis
• Mixed methods
• Visual tools
• Reports
• MAXdictio
• AI assist

There’s another menu below it.

There is this section, where we’ll see the data once we import it.

There is this section where we are going to see our codes once we try to get the codes.

Then there is this section where we are going to see the transcripts.  

I am going to have a single transcript that I want to analyze. I can drag and drop the transcript directly into the transcript section, and that is how I import that document.

Because this is a basic introduction, I want to take you back to the framework that I just highlighted.

If we are following this framework, the first thing that we will do is to familiarize ourselves with the data.

That means that I am going to read the transcript like a story to understand the context and the nuance of the responses.

 For example, let me read the transcript, and then we start coding.

These are the transcripts of focus group discussions among resident physicians assessing their perceptions of artificial intelligence, conducted from 2018 to 2019 at NewYork-Presbyterian Columbia University Medical Center.

The first question can be kind of hit or miss, depending on the field you are in, but do you guys have any exposure or experience with artificial intelligence like either in your careers your experiences now or even outside your other lives.

 Then the person responds, shortly like reading an article in the paper, that’s it.

Then the researcher ask them, do you guys use any AI as far as you are aware of in your current practice in these days in er?

And the interviewee says, I feel like I use Google Translate, but that’s it. No.

Research member: Do you guys know of anything coming down in the pipe in terms of like AI that people are trying to develop or look at in terms of emergency rooms?

 Then this person says: there is, uh. some people have been playing around with seeing if you can use AI for ESI and assigning ESI levels from triage based off like vital signs and age. I think it doesn’t , I seem to remember reading about something a while ago, I don’t think it performed much better than a human triages, but I think that’s one role that I’ve seen in terms of using it for ESI prediction research member. 

You guys haven’t heard of anything else, maybe. 

Speaker: no. 

How do you guys think AI could change emergency medicine in the future?

Speaker: So I think, yeah, I mean again, ESI would be useful I mean, basically triage and flow I think would be the most interesting ways of it affecting us, at least initially thinking about how we characterize patients and assess their severity and then also we move them through the department like who is capable for what Triage.

Basically, I’m just reading this transcript to familiarize myself.

I have gotten the essence that if I conducted this study, I would know what I did here through transcription, reading again, and cleaning up.

Once I familiarize with the data, I go to the second step suggested by Braun and Clarke, which is generating initial codes.

What are codes?

 A code is a label or an interpretive statement to any information that is important to our research question.

So by coding in MAXQDA, we are simply interpreting or tagging information that we believe is important.

For example, The first question can be kind of hit or miss, depending on the field you are in, but do you guys have any exposure or experience with artificial intelligence like either in your careers your experiences now or even outside your other lives.

 So the question is, do they have any experiences with artificial intelligence?

That’s the main question that this person is asking them.

Speaker says, short of like reading an article in the paper that is it.

Most of the time, I like using either color codes or questions as containers.

For this example, let us assume we go back to the original transcripts.

Let us assume we mark the first question as red.

The other question is green.

The other question is blue.

 The other question is marked as purple.

 The other question is gold and so on.

So why am I marking these questions?

Because I want to know where the codes came from.

 Let’s go back to MAXQDA and do some coding.

 Do you guys have any exposure or experience with artificial intelligence, like either in your careers your experiences now, or even outside your other lives?

Short of like reading an article in the paper that is it.

So this person says they have read about AI only they don’t have experience with it.

 Remember, this is between 2018 and 2019; AI was not as popular as it is today.

We can right-click and select New Code.

Then type, only read about AI.

 And the default color for that code is blue.

Remember, we tagged that question as red. 

So the color codes will show us where the codes came from originally, which will be important when combining codes to form themes, which is normally the other step of the Braun and Clarke.

 After we do all the coding, we have to combine our codes to form themes.

Then do you guys use any AI as far as you are aware of in your current practice these days In ER

 I feel like I use Google Translate, but that is it. NO.

 I think they Google Translate uses a form of AI to recognize language.

 We can just code this as, Google translates.

Remember the second question was green.

So I right click go to color and label it as green. 

Then do you guys know of anything like coming down In terms of like AI that people are trying to develop or look at, it in terms of emergency rooms 

There is, uh, some people have been playing around with seeing if you can use AI for ESI and assigning ESI levels, from triage based off like vital signs and age. I think it doesn’t, I seem to remember reading about something a while ago. I don’t think it performed much better than, than our human triagers. But I think that’s one role that I’ve seen in terms of using it for ESI prediction.

So if we highlight the whole paragraph, and code this as ESI prediction.

If you go back to our original transcript, you will see that the original question is light or sky blue, and the code already has the color blue so we leave it like that.

 You guys haven’t heard of anything else, maybe speak or how do you guys think AI could change emergency medicine in the future?

So, I think, yeah, I mean again, ESI would be useful. I mean basically triage and flow I think would be the most interesting ways of it affecting us, at least initially thinking about how we characterize patients and assess their severity and then also how we move them throughout the department. Like who’s like capable for vertical triage, who’s capable for labs and go, who has to come right back, who can wait in the waiting room. Um, how to decompress bottlenecks in terms of care. Like that’s probably the, I think we’re the highest yield is for us, since that’s so much of our efforts these days are like in operations is built around and like advanced modeling that could kind of think about how in real time to move patients around would probably really useful.

So this person says, how do they think AI will change medicine? Here they’re saying the use of AI to make the triage process better I think

I can highlight this and say improving triage.

 And we can assign gold color to this code.

Now you get the gist of it or the sense of coding.

Sometimes we can have more than two codes from the same question, or a question is asked from this transcript, and same question is asked to another participant in another transcript, that means the codes will have the same color code, which will help us easily combine these codes based on shared meaning to form themes when we use color codes or questions as containers.

So this is a basic introduction to using MAXQDA for conducting thematic analysis.

The main thing that I’ve said is you have to have an approach that you are using to conduct thematic analysis.

In this case, we have used the Braun and Clarke six-steps which include:

• Familiarizing with the data.
• Generating initial code.
• Developing themes.
• Reviewing themes.
• Defining and naming themes.
• Writing the report.

So this is what I would do, but for this article, I wanted to introduce coding and the use of MAXQDA, which is why I imported the transcript and showed you how to manage your codes.

After that, we combine the codes to form themes.