title: "Reading Scores and Class Size: A Classroom Study" author: A. Researcher bibliography: refs.bib abstract: | We compare reading scores in small and large classes in 80 schools. Pupils in small classes score 7.2 points higher on the end-of-year test.
Introduction
Class size is a long-standing question in education research [@alpha2015; @beta2020]. The design is described in @sec:design, the scores are in @tbl:scores and the classroom layout is in @fig:room.
Design {#sec:design}
We observed 80 schools during one school year, and 37.5% (30/80) of them had small classes. Schools were not assigned at random, so the comparison is descriptive. The prediction error is summarised by the root mean squared error (RMSE), and the RMSE is 4.1 points.
{#fig:room}
| School type | Small | Large | Gap |
|---|---|---|---|
| Urban | 71.2 | 64.0 | 7.2 |
| Rural | 69.5 | 63.1 | 6.4 |
: Mean reading scores by class size, in points out of 100. {#tbl:scores}
# TODO tune the model; contact @notakey
score = 1 + 1Conclusion
Small classes go together with higher reading scores in these schools. The data and the analysis code are available in a public archive, and the analysis was run in Python 3.12.