# Qualitative analysis and coding

Canonical page: https://www.questionpunk.com/support/qualitative-analysis

Code open-ended responses with AI assistance, build theme hierarchies, and measure intercoder reliability.

Time: 15-30 minutes

Level: Intermediate to Advanced

Audience: Researchers, Academics, Analysts

## Steps

### 1. Open qualitative analysis

Click **Qualitative analysis** in the sidebar's Analysis section to open a gallery of every study you have coded or imported, or pick **Qualitative coding** from the switcher on a survey's Results tab to open the same workspace scoped to that survey. You can also import open-ended data from outside QuestionPunk: upload a CSV, or connect a Qualtrics survey directly and pull in its text-entry questions.

### 2. Set up a research framework

In the **Codebook** tab, define your research questions and create an initial codebook. You can start with a blank framework or let AI suggest codes based on your data.

### 3. Run an AI coding session

In the **Coding** tab, start a coding session to have AI apply codes to your open-ended responses. The AI reads each response and assigns relevant codes from your codebook, suggesting new codes when it encounters themes not yet captured.

### 4. Review and refine codes

In the **Review** tab, review AI-assigned codes, accept or reject suggestions, and organize codes into theme hierarchies in the Codebook. Merge similar codes and split overly broad ones.

### 5. Track saturation and explore themes

Use the **Explore** tab to monitor coding saturation and see when new responses stop producing new themes. This helps determine when you have collected enough data for your research goals.

### 6. Measure intercoder reliability

In the **Calibration** tab, run intercoder reliability analysis to measure agreement between coders using Cohen's Kappa and Krippendorff's Alpha. This is essential for academic research requiring coding validity.

### 7. Export coded data

Use the **Export** tab to export your coded responses, theme hierarchies, and reliability metrics. Coded data can be included in CSV and Excel exports alongside response data.

## Details

Qualitative analysis in QuestionPunk bridges the gap between raw open-ended responses and structured research findings. AI-assisted coding accelerates the process while maintaining researcher control.

The workspace has eight tabs: **Overview**, **Data**, **Codebook**, **Calibration**, **Coding**, **Review**, **Explore**, and **Export**. It works both as a standalone, cross-survey workspace reached from the sidebar's **Analysis** section, and embedded in a native survey's Results tab alongside the Analysis sub-tab and Quantitative analysis.

The **Data** tab is not limited to native QuestionPunk surveys: import a CSV of open-ended responses, or connect directly to a Qualtrics survey and select which text-entry questions to bring in for coding.

Theme hierarchies let you organize codes into parent-child relationships, making it easy to analyze data at different levels of abstraction.

Saturation tracking shows you how quickly new themes emerge as more responses are coded, helping you make informed decisions about sample size adequacy.

Intercoder reliability metrics (Cohen's Kappa and Krippendorff's Alpha) provide quantitative measures of coding consistency, which is a requirement for publishable qualitative research.

Add reflexivity notes to document your analytical decisions and maintain an audit trail of your coding process.

## Related articles

- [AI insights and analysis](https://www.questionpunk.com/support/ai-insights.md)
- [Viewing the report](https://www.questionpunk.com/support/view-report.md)
- [Quantitative analysis](https://www.questionpunk.com/support/quantitative-analysis.md)
- [Exporting results](https://www.questionpunk.com/support/data-exports-exporting-results.md)