Evaluating Developmental Cognition Capabilities of LLMs

TL;DR

This paper introduces the Developmental Sentence Completion Test (DSCT) to evaluate the developmental cognition capabilities of large language models (LLMs) using a new, scalable assessment method based on Kegan's theory.

Contribution

It presents the DSCT as a novel, scalable instrument for assessing developmental signals in LLM responses, bridging a gap in existing evaluation methods.

Findings

LLMs recover simulated persona labels with high accuracy.

Human-LLM agreement on real responses is fair, with stronger within-neighborhood agreement.

Larger, newer models tend to generate higher developmental stage signals.

Abstract

Conversational AI is increasingly personalized around users' preferences, histories, goals, and knowledge, but much less around how users interpret and take up model outputs to construct and understand their reality. We draw on Robert Kegan's constructive-developmental theory as a complementary lens on this dimension. Existing methods for assessing developmental stage in the Keganian tradition rely either on expert interviews that do not scale or on sentence-completion instruments that are proprietary, lengthy, or invasive. To make this perspective tractable for LLM evaluation, we introduce the Developmental Sentence Completion Test (DSCT), a 20-item instrument designed to elicit developmental signal in self-administered text. Throughout, we treat the resulting labels as characterizations of stage-like structure in elicited responses, not as validated person-level developmental stage. We then ask how much of that signal can be recovered by LLMs across three elicited response regimes: simulated personas, real human respondents, and default model-generated answers. On simulated personas, top frontier models recover simulator-intended labels with high accuracy. On real human DSCT responses, human-LLM agreement is fair, with much stronger within-neighborhood than exact agreement. Finally, when LLMs answer DSCT prompts without persona-conditioning, their responses exhibit stable stage-like differences across model families, with larger and newer models tending to generate higher-rated text. These results suggest that stage-conditioned signal is cleaner in synthetic responses than in human-written DSCT text, and that the core constraint for stage-aware conversational AI is not classifier accuracy alone, but the availability of developmental signal from elicited text.