The first ESSA Tier II quasi-experimental study on Once — LXD Research finds a 0.46 standard deviation effect (p<0.001) for kindergarten students who received 80+ Once tutoring sessions.
Study at a glance
748 kindergarteners across 13 schools, all starting below the 50th percentile nationally.
Effect size by analysis
Hedges' g · scale to 0.50
All striving readers
N = 748 · p = .06 · trend-level
0.17
High attendance
N = 578 · p = .047 · statistically significant
0.20
80+ Once sessions
N = 568 · p < .001 · statistically significant
0.46
80+ session analysis
51.03
54.93
National standing
29%
Comparison
50%
Once, 80+ sessions
All students began below the 50th percentile. The difference in quartile movement was statistically significant at p < .001.
Source: LXD Research, 2024–25 Once Striving Readers Efficacy Study. Models controlled for baseline reading scores, attendance, student demographics, and school-level clustering.
We’re excited to share the results of the first quasi-experimental study (ESSA Tier II) on Once.
A quasi-experimental study estimates the causal impact of an intervention on a target population without using random assignment, but by matching treatment students to control students with similar demographics and similar BOY proficiency.
This study was conducted by LXD Research in a large southeastern school district on data from the 2024-25 school year. Fastbridge, by Renaissance Learning, is the nationally normed, computer-adaptive assessment that was used to measure growth.
The study had several significant findings. We’re most excited about the finding for kindergarten students who started the school year below the 50th percentile nationally on Fastbridge and received 80 or more Once tutoring sessions. This, we believe, measures the effect of the treatment on the treated, since prior correlational (ESSA Tier III) research has shown a disproportionate impact from Once on students who receive over 80 tutoring sessions compared to students who receive fewer than 80 tutoring sessions.
The effect that was observed here was 0.46 standard deviations with a p-value <0.001, meaning there is a less than 0.1% chance that the effect was the result of luck; the finding is statistically significant. The effect size is substantially larger than what other high-quality tutoring programs have reported in quasi-experimental studies and randomized controlled trials.
To be clear, this was a quasi-experimental study, not a randomized controlled trial, so the impact could have been confounded by other non-random factors that the researchers were not able to control for.
We have a separate randomized controlled trial in progress during school year 2025-26, and we look forward to sharing those results in the next several months.
Once — Quasi-Experimental Study Fastbridge Report 24-25, LXD Research (PDF)