Nadav Kunievsky, Pedro Pertusi
arXiv 6 Jan 2026 · Econometrics
arXiv:2601.03469 · PDF · DOI · OpenAlex · Extracted main text
Students from different socioeconomic backgrounds exhibit persistent gaps in test scores, gaps that can translate into unequal educational and labor-market outcomes later in life. In many assessments, performance reflects not only what students know, but also how effectively they can communicate that knowledge. This distinction is especially salient in writing assessments, where scores jointly reward the substance of students' ideas and the way those ideas are expressed. As a result, observed score gaps may conflate differences in underlying content with differences in expressive skill. A central question, therefore, is how much of the socioeconomic-status (SES) gap in scores is driven by differences in what students say versus how they say it. We study this question using a large corpus of persuasive essays written by U.S. middle- and high-school students. We introduce a new measurement strategy that separates content from style by leveraging large language models to generate multiple stylistic variants of each essay. These rewrites preserve the underlying arguments while systematically altering surface expression, creating a "generated panel" that introduces controlled within-essay variation in style. This approach allows us to decompose SES gaps in writing scores into contributions from content and style. We find an SES gap of 0.67 points on a 1-6 scale. Approximately 69% of the gap is attributable to differences in essay content quality, Style differences account for 26% of the gap, and differences in evaluation standards across SES groups account for the remaining 5%. These patterns seems stable across demographic subgroups and writing tasks. More broadly, our approach shows how large language models can be used to generate controlled variation in observational data, enabling researchers to isolate and quantify the contributions of otherwise entangled factors.
appendix boundary found by appendix_command · 73% of the source is main text. Read the extracted text to check this.
The works this paper leans on most, across its whole bibliography — not restricted to papers in our corpus. Ranked by composite intensity, which combines how often a work is mentioned, how many sections mention it, and how much of that falls in the main text rather than the appendix.
| Reference | Intensity | Mentions | Sections | Main text | |
|---|---|---|---|---|---|
| 1 | Potter, A., Shortt, M., Goldshtein, M., and Roscoe, R. D (2025) Assessing academic language in tenth grade essays using natural language processing | 0.585 | 3 | 1 | 100% |
| 2 | Crossley, S. A., Tian, Y., Baffour, P., Franklin, A., Benner, M., an… (2024) A large-scale corpus for assessing written argumentation: Persuade 2.0 | 0.511 | 2 | 1 | 100% |
| 3 | Kyle, K (2016) Measuring syntactic development in L2 writing: Fine-grained indices of syntactic complexity and usage-based indices of syntactic… | 0.511 | 2 | 1 | 100% |
| 4 | Kyle, K., Crossley, S. A., and Jarvis, S (2021) Assessing the validity of lexical diversity using direct judgements | 0.511 | 2 | 1 | 100% |
| 5 | Blinder, A. S (1973) Wage discrimination: Reduced form and structural estimates | 0.405 | 1 | 1 | 100% |
| 6 | Kitagawa, E. M (1955) Components of a difference between two rates | 0.405 | 1 | 1 | 100% |
| 7 | Oaxaca, R (1973) Male–female wage differentials in urban labor markets | 0.405 | 1 | 1 | 100% |
| 8 | Arnold, D., Dobbie, W., and Hull, P (2022) Measuring racial discrimination in bail decisions | 0.405 | 1 | 1 | 100% |
| 9 | Bohren, J. A., Hull, P., and Imas, A (2022) Systemic discrimination: Theory and measurement | 0.405 | 1 | 1 | 100% |
| 10 | Clark, K., Khandelwal, U., Levy, O., and Manning, C. D (2019) What does BERT look at? an analysis of BERT's attention | 0.405 | 1 | 1 | 100% |
Showing the top 10 of 35 scored citations.