Data segmentation is the process of dividing data into meaningful units for coding and further analysis. Segmentation helps define the boundaries of observations, interactions, or discourse that will be coded and analyzed. The way data are segmented influences the patterns and relationships that can be identified in a QE study.
Here you’ll find information and resources on preparing your qualitative and multimodal data for QE analysis, including:
| Data segmentation | ||
|---|---|---|
| Parsing the Continuum: Manual Segmentation of Monologic Data | Zörgő, S., & Brohinsky, J. (2023) | If you are unsure how to break your qualitative data into meaningful parts for analysis, this guide offers practical tips for segmentation and shows why these decisions are just as important as your coding choices. |
| Choosing Units of Analysis in Temporal Discourse | Barany, A., Philips, M., Kawakubo, A. J. T., & Oshima, J. (2022) | This study explores how different ways of slicing your data into chronological time units affect your final ENA models, testing a computer-assisted tool (KBDeX) against simple shortcuts to see if they can match the accuracy of manual grouping. |
| Segmentation and Code Co-occurrence Accumulation: Operationalizing Relational Context with Stanza Windows | Zörgő, S. (2023) | To help you decide how to best divide your data, this paper explores three different methods for segmentation (chunking text) and demonstrates how each choice directly affects the visual patterns you see when using Epistemic Network Analysis (ENA). |
| In Search of Conversational Grain Size: Modeling Semantic Structure Using Moving Stanza Windows | Siebert-Evenstone, A. L., Arastoopour Irgens, G., Collier, W., Swiecki, Z., Ruis, A. R., & Shaffer, D. W. (2017) | This paper examines the affordances of temporal segmentation (stanza) for modelling connections in discourse. |
| Exploring the Effects of Segmentation on Semi-structured Interview Data with Epistemic Network Analysis | Zörgő, S., Swiecki, Z., & Ruis, A. R. (2021) | This paper uses interview data to explore how the way you segment your text changes the patterns you see at both the individual and whole group level when using Epistemic Network Analysis (ENA). |
| Data coding | ||
|---|---|---|
| How We Code | Shaffer, D. W., & Ruis, A. R. (2021) | This paper provides guidelines for good coding practice and highlights some of the issues and key questions that quantitative ethnographers and other researchers should consider when coding data. |
| Analyzing Interview Data: The Development and Evolution of a Coding System | Weston, C., Gandell, T., Beauchamp, J., McAlpine, L., Wiseman, C., & Beauchamp, C. (2001) | This paper walks through how a research team developed their qualitative coding system, showing that creating codes together is actually a crucial part of the analysis itself. |
| Approaches to Code Selection for Epistemic Networks | Árva, D., Jeney, A., Dunai, D., Major, D., Cseh, A., & Zörgő, S. (2023) | To help you manage projects with a large number of codes, this paper outlines three practical approaches for selecting the most relevant codes so your final ENA models remain clear and easy to interpret. |
| A Lightweight Interactive Regular Expression Generator for Qualitative Coding in Quantitative Ethnography | Cai, Z., Marquart, C., Eagan, B., Xiao, Y., & Shaffer, D. W. (2023) | This paper introduces an interactive tool that helps researchers create transparent, automated text-coding rules with minimal pre-coded data, making large qualitative datasets easier to analyze. |
| From nCoder to ChatGPT: From Automated Coding to Refining Human Coding | Zambrano, A. F., Liu, X., Barany, A., Baker, R. S., Kim, J., & Nasir, N. (2023) | This paper finds that nCoder works well for straightforward keyword-based coding, while ChatGPT can recognize more varied language and explain its decisions to help researchers clarify codes, although human judgment remains essential for complex concepts. |
| CritQE: A Pathway Toward Convergence and Equity | Garcia, N. M., Onyewuenyi, A. C., & Shaffer, D. W. (2025) | To help researchers center equity in their work, this paper introduces Critical Quantitative Ethnography (CritQE), a framework combining QE with social theory, and demonstrates how multivocal coding can be engaged as a technique that uses the disagreements between your coders to uncover diverse insights instead of just viewing them as mistakes. |
| Comparing Human and LLM Ordered Coding of Qualitative Data: How Coding Differences Cascade Through Temporal Analysis | Misiejuk, K., Lopez-Pernas, S., Oliveira, E. A., Eagan, B., & Saqr, M. (2026) | This study finds that large language models do not yet reliably reproduce the order and structure of human qualitative coding, and that their coding differences can lead to substantially different representations and interpretations of learning processes. |
| Reliability | ||
|---|---|---|
| Can We Rely on IRR? Testing the Assumptions of Inter-Rater Reliability | Eagan, B. R., Rogers, B., Serlin, R., Ruis, A. R., Arastoopour Irgens, G., & Shaffer, D. W. (2017) | This research explains why you must be careful when testing if your coders agree, revealing through simulations that the most widely used calculation for checking this agreement frequently produces errors. |
| Testing the Reliability of Inter-Rater Reliability | Eagan, B., Brohinsky, J., Wang, J., & Shaffer, D. W. (2020) | This paper shows that common methods for determining whether coders agree can mistakenly label unreliable coding as reliable, regardless of the agreement measure used, and calls for stronger methods that control such errors. |
| Neural Recall Network: A Neural Network Solution to Low Recall Problem in Regex-based Qualitative Coding | Cai, Z., Marquart, C., & Shaffer, D. W. (2022) | This paper shows how neural networks can help researchers identify examples missed by keyword-based coding, making it faster to develop more complete, transparent, and scalable tools for analyzing qualitative text. |
| LSTM Neural Network Assisted Regex Development for Qualitative Coding | Cai, Z., Eagan, B., Marquart, C., & Shaffer, D. W. (2023) | This paper shows that using a neural network to flag text likely missed by keyword-based coding can help researchers improve their coding rules while reducing the amount of manual coding required by approximately 50%–63%. |
| Does Active Learning Reduce Human Coding? A Systematic Comparison of Neural Network with nCoder | Choi, J., Ruis, A. R., Cai, Z., Eagan, B., & Shaffer, D. W. (2023) | This paper finds that nCoder’s active-learning approach can produce highly accurate automated qualitative coding with significantly less human-coded training data than a traditional neural network. |
| Interrater Reliability: The Kappa Statistic | McHugh, M. L. (2012) | To ensure your data is accurately coded, this paper explains how to measure whether different team members agree on their coding choices—known as interrater reliability—and helps you determine the best statistical standard for your specific project. |
| From nCoder to ChatGPT: From Automated Coding to Refining Human Coding | Zambrano, A.F., Liu, X., Barany, A., Baker, R.S., Kim, J., & Nasir, N. (2023). | This paper finds that nCoder works well for straightforward keyword-based coding, while ChatGPT can recognize more varied language and explain its decisions to help researchers clarify codes—although human judgment remains essential for complex concepts |
| Perfect Sampling | Shaffer, D. W., & Cai, Z. (2024) | This paper introduces a method for calculating the smallest sample on which two coders must agree perfectly to demonstrate that their overall coding agreement reliably exceeds a chosen standard. |
| ChatGPT for Education Research: Exploring the Potential of Large Language Models for Qualitative Codebook Development | Barany, A., Nasir, N., Porter, C., Zambrano, A. F., Andres, A. L., Bright, D., Shah, M., Gao, S., Zhang, J., Mehta, S., Choi, J., Giordano, C., & Baker, R. S. (2024) | This study finds that ChatGPT can help researchers develop qualitative codebooks more quickly and reliably when used in partnership with people, but fully automated codebook development cannot replace human judgment. |
| Subgroup Fairness in Multilingual Text Classification | Ruis, A. R., Cai, Z., & Shaffer, D. W. (2025) | This paper examines whether automated text-coding systems interpret equivalent content fairly across languages and introduces a technique that significantly reduces language-related coding errors. |
| Comparing Human and LLM Ordered Coding of Qualitative Data: How Coding Differences Cascade through Temporal Analysis | Misiejuk, K., Lopez-Pernas, S., Oliveira, E.A., Eagan, B., & Saqr, M. (2026). | This study finds that large language models do not yet reliably reproduce the order and structure of human qualitative coding, and that their coding differences can lead to substantially different representations and interpretations of learning processes. |
| Additional resources | |
|---|---|
| Coded Datasets | Examples of coded datasets to download and use to practice QE modeling techniques, such as ENA, ONA, and TMA. |
| QE Data Challenge | Access sample datasets from previous QE Data Challenges. |
Browse our Zotero library for a comprehensive, continually updated collection of Quantitative Ethnography research.