Inductive Thematic Analysis in ATLAS.ti: Saldana’s Method

If you want to conduct inductive thematic analysis in ATLAS.ti using Saldana’s method, you’re in the right place.

In this step-by-step guide, I’ll walk you through all four steps — from generating your first codes to connecting finished themes to your research questions — using real interview data and screenshots from ATLAS.ti.

However, conducting inductive thematic analysis is one of the most popular approaches to analyzing qualitative data.

Quick summary

Quick summary: Saldana’s approach to inductive thematic analysis involves four steps: (1) identifying and categorizing codes, (2) developing high-level categories, (3) generating themes through analytic memoing, and (4) connecting themes to your research questions.

There are two main frameworks researchers use for thematic analysis. The first is the Braun and Clarke six-step framework.

The second is Saldana’s approach — which is what we’ll follow in this tutorial. If you’re new to ATLAS.ti, I’d recommend starting with how to do qualitative analysis of interviews with ATLAS.ti before continuing here.

My name is Bernard Mugo. Over the past three years, I’ve helped more than 300 PhD students analyze qualitative data and complete their theses. This guide is based on that hands-on experience.

What Is Inductive Thematic Analysis? (Saldana’s Approach)

Inductive thematic analysis is an approach to qualitative data analysis where themes emerge directly from the data — rather than being defined in advance.

Saldana’s method structures this process into four clear steps, making it especially practical for PhD students working with interview transcripts.

There are two common approaches of doing inductive thematic analysis, which include the Braun and Clarke six-step framework.

Here are the four steps we’ll follow in ATLAS.ti:

1. Identifying and categorizing codes
2. Developing high-level categories
3. Generating themes through analytic memoing
4. Connecting themes to research questions

Let’s work through each one.

Step 1 — Identifying and Categorizing Codes in ATLAS.ti

What Is a Code?

Before we open ATLAS.ti, let’s define what a code actually is. A code is a label or interpretive statement assigned to any piece of data that is important to your research questions.

To make this concrete, here’s an excerpt from an interview transcript:

“Mrs. Jackson rises from her desk and announces: ‘Okay, you guys — let’s get lined up for lunch.’ Row one, five children seated in the first row of desks rise and walk to the classroom door. Some of the seated children talk to each other. Mrs. Jackson looks at them and says: ‘No talking — save it for the cafeteria.’ Row two, five children rise and walk to join the line.”

From this paragraph, we can generate three codes:

• Lining up for lunch (sentence 1 and 3)
• Managing behavior (sentence 2)
• Lining up for lunch (sentence 3 — same code applied again)

Now let’s apply this logic inside ATLAS.ti.

How to Create and Apply Codes in ATLAS.ti

Open ATLAS.ti and create a new project — I’ll call mine “Saldana Approach.” In the Documents section, drag and drop your interview transcripts. I’m working with two transcripts here.

For one of my transcripts, I used color codes to mark which question each response belongs to — this makes categorizing codes much easier later.

Coding Transcript 1:

For the first question — “In your opinion, what kind of teaching and learning challenges do students in your class experience?” — the response mentions a broken air conditioning unit and poor ventilation.

I highlight that section, right-click, select Apply Codes, and create the code: Infrastructural challenges.

I assign this code the color red to mark it as coming from question one.

Coding Transcript 2:

When I move to the second transcript, I look for the same codes.

Where a response matches an existing code — like “Infrastructural challenges” — I apply that existing code rather than creating a new one. Double-clicking the code shows me it now has two quotes from two different participants, which strengthens the evidence for that code.

This is the power of ATLAS.ti for Saldana’s first step: you build a shared codebook across all your transcripts simultaneously.

Step 2 — Developing High-Level Categories in ATLAS.ti

With all codes generated, the second step is grouping them into high-level categories based on meaning.

I create a folder for each interview question and drag the corresponding color-coded codes into it:

• Category A — Teaching and learning challenges (red codes)
• Category B — Challenges experienced by students (blue codes)
• Category C — Teaching strategies (purple codes)
• Category D — Effectiveness of teaching strategies (green codes)
• Category E — Effective approaches of teaching students (remaining codes)

At this stage, you’ll notice some codes might seem to fit multiple categories. Don’t worry — we’ll refine this in Step 3.

Step 3 — Generating Themes Through Analytic Memoing

Now we move from categories to themes, which is where the real analytical work happens. In ATLAS.ti, I rename each category folder to reflect its theme and add a written memo to describe what it represents.

For example:

• Theme A — Teaching and learning challenges: “This theme represents the different teaching and learning challenges that students and teachers experience in the community college.”
• Theme B — Challenges experienced by students: “This theme highlights all the different challenges experienced by students in the community college.”
• Theme C — Teaching strategies: “This theme represents the different strategies applied by teachers to impart knowledge among students.”

As I write these memos, I start to see that some of my five original categories overlap. I review each code and ask: does it really belong here, or does it fit better elsewhere?

• High levels of absenteeism → moved to Theme B (it’s a student challenge, not a teaching strategy)
• Collaborative peer teaching → moved to Theme C (it’s a teaching strategy)
• Lack of sufficient time to finish the syllabus → moved to Theme A (it’s a teaching and learning challenge)
• Encouraging active engagement → merged with Promotes active learning among students (they mean the same thing)

After this review, my five categories have condensed into three clean, well-defined themes. That’s Saldana’s method working exactly as intended.

Step 4 — Connecting Themes to Your Research Questions

The final step is applying your themes to your original research questions. I import both research questions directly into ATLAS.ti as documents:

• RQ1: What are the main challenges teachers experience when working in a community college?
• RQ2: What are the different teaching strategies employed by teachers in a community college?

Then I drag the relevant themes into each research question:

• RQ1 ← Theme A (Teaching and learning challenges) + Theme B (Challenges experienced by students)
• RQ2 ← Theme C (Teaching strategies)

The final structure shows two themes answering RQ1 and one theme answering RQ2 — a clean, defensible analytical outcome that’s directly grounded in your data.

Final Thoughts

That’s how to conduct inductive thematic analysis in ATLAS.ti using Saldana’s four-step method. To recap: you start by generating codes from your transcripts, group them into high-level categories, refine those into themes through analytic memoing, and finally connect those themes to your research questions.

If you want to apply the same Saldana method in a different tool, check out:

• Inductive thematic analysis using NVivo (Saldana method)
• Saldana’s method of inductive thematic analysis using MAXQDA

And if you’re working with Saldana’s book — The Coding Manual for Qualitative Researchers — this tutorial follows his framework closely.

Need hands-on help with your own qualitative analysis? Reach out here — I work with PhD students one-on-one to get their data analyzed and findings written up.