Deductive vs. inductive approaches in qualitative research

Who this article is for

This article is for researchers, students, instructors, evaluators, and qualitative analysts who want to apply deductive or inductive approaches in ATLAS.ti.

It is especially useful if you are working with:

• interviews
• focus groups
• field notes
• literature reviews
• open-ended survey responses
• exploratory qualitative research
• theory-testing research
• thematic analysis
• grounded theory workflows

Deductive and inductive approaches are not mutually exclusive. Many qualitative researchers combine both approaches by starting with predefined concepts while also allowing new ideas and patterns to emerge from the data during analysis.

What are deductive and inductive approaches in qualitative research?

There are two common approaches to qualitative data analysis: deductive and inductive approaches.

Deductive analysis works from broader theories or concepts toward specific observations in the data. This is sometimes called a top-down approach. Researchers begin with predefined codes, concepts, or theoretical frameworks and apply them to their materials.

Inductive analysis works in the opposite direction, moving from specific observations toward broader concepts and theories. This is often called a bottom-up approach. Researchers develop codes, categories, and themes directly from the data as they analyze it.

Both approaches are valid and can be used in ATLAS.ti. Many qualitative projects use a hybrid workflow that combines predefined codes with emergent coding throughout the research process.

Use a deductive approach in ATLAS.ti Web

A deductive approach begins with predefined theories, concepts, categories, or codes. Researchers apply these existing ideas to the data to evaluate, refine, or extend a theoretical framework.

Step 1: Create a project and add documents

1. Log in to ATLAS.ti Web.
2. Create a new project.
3. Open the Document Manager.
4. Click Add Documents.
5. Upload your transcripts, survey responses, field notes, PDFs, or other qualitative data.

Step 2: Create or prepare a deductive code system

Researchers using deductive analysis often begin with a predefined set of codes based on:

• a theoretical framework
• research questions
• literature reviews
• interview guides
• prior studies
• an existing codebook

In ATLAS.ti Web, the Code Manager lets users view, edit, delete, comment on, group, merge, and export codes. It also displays groundedness, comments, and creation dates. 

Option 1: Create the code system manually in ATLAS.ti Web

1. Open the Code Manager by clicking the codes icon in the left-hand menu.
2. Create each predefined code manually.
3. Add the code name.
4. Repeat this process until your deductive code list is complete.
5. Add code comments to define each code.
6. Organize related codes into groups, categories, or subcodes as needed.

Option 2: Prepare the code system in ATLAS.ti Desktop

Use this option if you already have a larger codebook that you do not want to recreate manually in ATLAS.ti Web.

1. Export your ATLAS.ti Web project.
2. Open the project in ATLAS.ti Desktop.
3. Import or create the codebook in ATLAS.ti Desktop.
4. Check that the code names, comments, and structure are correct.
5. Export the updated project from ATLAS.ti Desktop.
6. Import the updated project back into ATLAS.ti Web.

Step 3: Add code comments for deductive definitions

Use code comments to document the meaning of each predefined code.

1. Go to the Code Manager.
2. Hover over the code you want to edit.
3. Click Details to open the sidebar.
4. Click in the comment space.
5. Write the code definition.
6. Include when the code should and should not be applied.
7. Repeat this process for each deductive code.

Use code comments for operational definitions, inclusion criteria, exclusion criteria, and examples. ATLAS.ti Web also allows users to download a codebook containing codes and their definitions.

Step 4: Apply predefined codes to the data

1. Open a document.
2. Read the text carefully.
3. Highlight a relevant segment.
4. Use the coding window to search for an existing code.
5. Select the predefined code.
6. Apply the code to the quotation.
7. Continue reviewing the document systematically.

Deductive coding often focuses on identifying evidence that supports, expands, or challenges an existing theory or framework.

Step 5: Rename codes in the Code Manager

Rename codes if the wording is unclear or if your codebook changes during analysis.

1. Go to the Code Manager.
2. Hover over the code you want to rename.
3. Click Details to open the sidebar.
4. Click the code name at the top of the sidebar.
5. Type the new code name.
6. Review the updated code name in the Code Manager.

Step 5: Rename codes in the Code Manager

Rename codes if the wording is unclear or if your codebook changes during analysis.

1. Go to the Code Manager.
2. Hover over the code you want to rename.
3. Click Details to open the sidebar.
4. Click the code name at the top of the sidebar.
5. Type the new code name.
6. Review the updated code name in the Code Manager.

Step 6: Change code colors

Use colors if you want to visually distinguish code types, categories, or phases of analysis.

1. Go to the Code Manager.
2. Click the code you want to edit.
3. Click the color circle next to the code name.
4. Choose the desired color.
5. Repeat this process for other codes if needed.

Step 8: Organize codes into code groups

Use code groups to organize and filter codes during analysis.

1. Go to the Code Manager.
2. Hover over a code.
3. Click Details to open the sidebar.
4. Click Edit groups.
5. Select an existing group or create a new group.
6. To manage groups, click the gear icon or Manage Groups.
7. Rename, edit, or delete groups as needed.

To add multiple codes to a group:

1. Go to the Code Manager.
2. Tick the boxes to the left of the codes you want to group.
3. Click the three dots at the bottom of the screen.
4. Select Add to group.
5. Choose the desired code group.

Step 9: Merge codes

Merge codes when two or more codes represent the same concept.

1. Go to the Code Manager.
2. Tick the boxes to the left of the codes you want to merge.
3. Click the three dots at the bottom of the screen.
4. Select Merge codes.
5. Enter the name of the new merged code.
6. Review the merged code in the Code Manager.

When codes are merged in ATLAS.ti Web, the audit trail of the merged codes is saved in the comment field. 

Step 10: Export codes

Export your code list if you want to save or share the codebook.

1. Go to the Code Manager.
2. Click the three dots in the top-right corner.
3. Select Download Codes as QDC or Download Codes as Excel.
4. Save the file in your preferred location.

Use an inductive approach in ATLAS.ti Web

An inductive approach develops codes, categories, and themes directly from the data. Instead of starting with a predefined theory, researchers identify patterns and concepts as they analyze their materials.

Step 1: Create a project and add documents

1. Log in to ATLAS.ti Web.
2. Create a new project.
3. Open the Document Manager.
4. Click Add Documents.
5. Upload your qualitative data.

Step 2: Start open coding

1. Open a document.
2. Read the data closely.
3. Highlight a meaningful segment of text.
4. Create a new code that reflects the meaning of the segment.
5. Apply the code to the quotation.
6. Continue coding throughout the document.

In inductive analysis, codes often begin as short descriptive labels close to the participants’ language or experiences.

Step 3: Compare quotations and refine codes

1. Open the Quotation Manager.
2. Review quotations linked to the same code.
3. Compare similarities and differences across quotations.
4. Check whether the quotations represent the same concept.
5. Rename unclear codes in the Code Manager.
6. Merge overlapping codes if needed.
7. Add or revise code comments.

This process is often called constant comparison.

Step 4: Organize emergent codes

1. Open the Code Manager.
2. Review the list of codes created during open coding.
3. Identify codes that describe similar concepts.
4. Use code groups to organize related codes.
5. Merge duplicate or overlapping codes.
6. Rename codes that are too vague.
7. Add code comments to explain what each code means.

Step 5: Use AI Coding and Intentional AI Coding carefully

AI Coding and Intentional AI Coding can support exploratory and inductive workflows by helping researchers identify possible patterns, themes, and relationships in the data. ATLAS.ti describes Intentional AI Coding as a way to create tailored codes aligned with specific research needs, while AI Code Suggestions can help users explore data from new perspectives. 

1. Open a document or select multiple documents.
2. Open AI Coding or Intentional AI Coding.
3. Review the suggested quotations, categories, and codes.
4. Select the results you want to apply.
5. Apply the selected results.
6. Review the AI-generated codes in the Code Manager.
7. Rename, merge, group, or delete codes as needed.
8. Add your own code comments and analytic interpretation.

AI Coding can help accelerate exploratory analysis, especially with larger datasets. However, qualitative interpretation and theory development remain researcher-driven processes.

Use deductive and inductive approaches in ATLAS.ti Desktop

ATLAS.ti Desktop includes ATLAS.ti Windows and ATLAS.ti Mac. Both versions support deductive, inductive, and hybrid workflows using codes, quotations, code comments, code groups, reports, networks, and analysis tools.

The Code Manager in both Windows and Mac is used to create and modify codes, code data segments, retrieve coded segments, organize codes, merge and split codes, filter codes, review codes in networks, and create reports

Step 1: Create a project and add documents

In ATLAS.ti Windows

1. Open ATLAS.ti Windows.
2. Click New Project.
3. Name your project.
4. Add your documents.
5. Import transcripts, PDFs, field notes, documents, or other qualitative data.

In ATLAS.ti Mac

1. Open ATLAS.ti Mac.
2. Click  New Project.
3. Name your project.
4. Add your documents.
5. Import transcripts, PDFs, field notes, documents, or other qualitative data.

Step 2: Create or import a deductive code system

Use this step if you are starting with predefined codes from a theory, framework, literature review, interview guide, or existing codebook.

In ATLAS.ti Windows

1. Open the Code Manager.
2. Create new codes manually, or import an existing code list or codebook if available.
3. Add each predefined code to the project.
4. Select a code.
5. Use Edit Comment to add a definition or coding rule.
6. Repeat this process for each code.
7. Use Rename if a code name needs to be adjusted.
8. Organize related codes into folders, groups, categories, or subcodes.

In ATLAS.ti Mac

1. Open the Code Manager from the toolbar or Code menu.
2. Create new codes manually, or import an existing code list or codebook if available.
3. Add each predefined code to the project.
4. Select a code.
5. Add the code definition in the comment area or inspector.
6. Repeat this process for each code.
7. Rename unclear codes if needed.
8. Organize related codes into groups, categories, or subcodes.

Step 3: Start inductive coding from the data

Use this step if you want to create codes while reading and interpreting your data.

In ATLAS.ti Windows

1. Open a document.
2. Read the text closely.
3. Highlight a meaningful segment.
4. Right-click the highlighted segment.
5. Select the coding option.
6. Create a new code that reflects the meaning of the segment.
7. Apply the code.
8. Continue coding line by line, paragraph by paragraph, or section by section.

In ATLAS.ti Mac

1. Open a document.
2. Read the text closely.
3. Highlight a meaningful segment.
4. Right-click the selected segment or use the coding toolbar.
5. Create a new code that reflects the meaning of the segment.
6. Apply the code.
7. Continue coding throughout the document.

Step 4: Apply existing codes to quotations

Use this step when applying deductive codes or reusing inductive codes that already exist.

In ATLAS.ti Windows

1. Open a document.
2. Highlight a relevant segment.
3. Right-click the highlighted segment.
4. Choose the option to apply codes.
5. Select one or more existing codes.
6. Apply the selected codes to the quotation.

In ATLAS.ti Mac

1. Open a document.
2. Highlight a relevant segment.
3. Right-click the selection or use the coding toolbar.
4. Select one or more existing codes.
5. Apply the selected codes to the quotation.

You can also use drag and drop from the Code Manager to code a highlighted data segment.

Step 5: Review quotations connected to a code

Use this step to compare coded data and refine your interpretation.

In ATLAS.ti Windows

1. Open the Code Manager.
2. Double-click a code.
3. Review the linked quotations in the Quotation Reader.
4. Browse through the quotations.
5. Open quotations in context if needed.
6. Compare whether the quotations reflect the same concept.

In ATLAS.ti Mac

1. Open the Code Manager.
2. Double-click a code.
3. Review the linked quotations in the Quotation Reader.
4. Browse through the quotations.
5. Open quotations in context if needed.
6. Compare similarities and differences across quotations.

In both Windows and Mac, double-clicking a code opens the linked quotations in the Quotation Reader.

Step 6: Rename codes

Rename codes when the wording is unclear, too broad, or no longer reflects your interpretation.

In ATLAS.ti Windows

1. Open the Code Manager.
2. Select the code you want to rename.
3. Use Rename in the Code Manager ribbon or context menu.
4. Enter the new code name.
5. Review the updated code name in the code list.

In ATLAS.ti Mac

1. Open the Code Manager.
2. Select the code you want to rename.
3. Right-click the code or use the available rename option.
4. Enter the new code name.
5. Review the updated code name in the code list.

Step 7: Merge overlapping codes

Merge codes when two or more codes refer to the same concept.

In ATLAS.ti Windows

1. Open the Code Manager.
2. Select two or more codes.
3. Open the Tools or context menu.
4. Select Merge Codes.
5. Choose or enter the final merged code name.
6. Review the merged code and its linked quotations.

The Windows Code Manager includes a Merge Codes option. When codes with comments are merged, comments are added to the merged code and information about the merged codes is saved as an audit trail. 

In ATLAS.ti Mac

1. Open the Code Manager.
2. Select two or more codes.
3. Open the context menu or available code action menu.
4. Select the merge option.
5. Choose or enter the final merged code name.
6. Review the merged code and its linked quotations.

Step 8: Organize codes into groups, categories, folders, or subcodes

Use this step to move from a long code list toward a clearer analytic structure.

In ATLAS.ti Windows

1. Open the Code Manager.
2. Review the full code list.
3. Select related codes.
4. Create a code group, folder, category, or subcode structure.
5. Move related codes into the structure.
6. Review whether the structure reflects your deductive framework or inductive findings.

In ATLAS.ti Mac

1. Open the Code Manager.
2. Review the full code list.
3. Select related codes.
4. Create a code group, category, or subcode structure.
5. Move related codes into the structure.
6. Review whether the structure reflects your deductive framework or inductive findings.

Step 9: Use analysis tools to compare patterns

Use analysis tools after you have coded enough data to compare patterns.

In ATLAS.ti Windows

1. Open the Code Manager.
2. Select one or more codes.
3. Use analysis options such as Code Co-occurrence or Code-Document Analysis.
4. Review which codes appear together.
5. Compare how codes appear across documents or document groups.
6. Open the quotations behind the results to interpret the pattern.

In ATLAS.ti Mac

1. Open the Code Manager.
2. Select one or more codes.
3. Open the relevant analysis tool, such as Code Co-occurrence or Code-Document Analysis.
4. Review which codes appear together.
5. Compare how codes appear across documents or document groups.
6. Open the quotations behind the results to interpret the pattern.

Step 10: Use networks to develop relationships between concepts

Networks can help you visualize relationships between deductive concepts, inductive themes, categories, quotations, and memos.

In ATLAS.ti Windows

1. Select codes, quotations, categories, or memos.
2. Open them in a network.
3. Add related concepts to the network.
4. Link concepts that appear connected.
5. Name or describe the relationships.
6. Use the network to review your developing interpretation.

In ATLAS.ti Mac

1. Select codes, quotations, categories, or memos.
2. Open them in a network.
3. Add related concepts to the network.
4. Link concepts that appear connected.
5. Name or describe the relationships.
6. Use the network to review your developing interpretation.

Step 11: Export reports or codebooks

Use reports or exported codebooks to review your analysis, share your coding structure, or document your methodology.

In ATLAS.ti Windows

1. Open the Code Manager, Quotation Manager, or another manager.
2. Select the codes, quotations, or items you want to include.
3. Create a report or export.
4. Save the file.

In ATLAS.ti Mac

1. Open the Code Manager, Quotation Manager, or another manager.
2. Select the codes, quotations, or items you want to include.
3. Create a report or export.
4. Save the file.

Combine deductive and inductive approaches

Many qualitative projects combine both approaches.

For example, researchers may:

• begin with a prepared codebook based on prior literature
• add new emergent codes during analysis
• refine categories iteratively
• revise assumptions as new insights appear
• use AI Coding or Intentional AI Coding to explore possible patterns
• manually review and interpret all coding results

This combined workflow is sometimes called a hybrid approach.

Step 1: Start with your initial framework

1. Create or import your predefined deductive codes.
2. Add code comments and definitions.
3. Apply the codes to early documents.
4. Review whether the codes fit the data.

Step 2: Add emergent codes

1. Continue reading the data closely.
2. Create new codes when important ideas do not fit the existing framework.
3. Add comments to explain the new codes.
4. Compare new codes with existing codes.

Step 3: Refine the full code system

1. Review all codes in the Code Manager.
2. Rename unclear codes.
3. Merge overlapping codes.
4. Organize related codes into groups, categories, folders, or subcodes.
5. Review linked quotations.
6. Update your code comments and memos.

Common issues and mistakes

• Creating overly broad deductive codes

  • Broad codes can make analysis difficult to interpret later. Create clear code definitions and refine large codes into smaller subcodes when necessary.

    Creating too many disconnected inductive codes

  • Inductive coding often starts with many codes, but these should gradually be organized into clearer categories and themes.

    Treating AI-generated codes as final analysis

  • AI Coding can support exploration and organization, but researchers should always review, refine, and interpret AI-generated results manually.

    Mixing categories and code groups unintentionally

  • Categories and subcodes create hierarchical relationships between concepts, while code groups are mainly used for organization and filtering.

    Skipping memo writing and code comments

  • Code comments and memos help document analytic decisions, evolving interpretations, and theoretical insights throughout the project.

    Avoiding revision of the coding framework

  • Qualitative analysis is iterative. Researchers often revise code definitions, categories, and interpretations as their understanding develops.

When to contact support

Contact ATLAS.ti Support if:

• documents cannot be imported
• codes or categories are not displaying correctly
• quotations do not appear as expected
• AI Coding tools fail to generate results
• categories, groups, folders, or subcodes cannot be created
• codes cannot be renamed, merged, deleted, or exported
• reports or exports fail
• features behave differently from the documentation

When contacting support, include:

• your platform: Web, Windows, or Mac
• your ATLAS.ti version if using Desktop
• screenshots or error messages
• a short description of your workflow
• whether you are using deductive, inductive, hybrid, or AI-supported workflows