Bridging Human Interpretation and Machine Representation: A Landscape of Qualitative Data Analysis in the LLM Era

TL;DR

This paper introduces a landscape framework to categorize and analyze how large language models support qualitative data analysis, highlighting current limitations and proposing an agenda for more explicit interpretive and dynamic modeling.

Contribution

It presents a novel 4x4 landscape framework to classify LLM outputs in qualitative research and identifies gaps in interpretive and dynamic modeling capabilities.

Findings

Current LLM outputs are skewed towards low-level, non-interpretive representations.

Few systems reliably perform interpretive or theoretical inference.

The landscape reveals significant gaps in dynamic and causal modeling in LLM-based qualitative analysis.

Abstract

LLMs are increasingly used to support qualitative research, yet existing systems produce outputs that vary widely--from trace-faithful summaries to theory-mediated explanations and system models. To make these differences explicit, we introduce a 4$\times$4 landscape crossing four levels of meaning-making (descriptive, categorical, interpretive, theoretical) with four levels of modeling (static structure, stages/timelines, causal pathways, feedback dynamics). Applying the landscape to prior LLM-based automation highlights a strong skew toward low-level meaning and low-commitment representations, with few reliable attempts at interpretive/theoretical inference or dynamical modeling. Based on the revealed gap, we outline an agenda for applying and building LLM-systems that make their interpretive and modeling commitments explicit, selectable, and governable.