Penda’s Impact Validated by Two Independent Studies Showing Major Student Growth

Multi-District, Meta-Analysis of a Science Intervention

Abstract

This meta-analysis synthesized 13 independent school-based data samples (N = 33,306 students) that compared the effects of the Penda Learning science platform with those of control cohorts, using random-effects modeling. Drawing data from multiple school districts, this analysis synthesized evidence from studies that meet the definition of Tier I–III evidence requirements outlined in the Every Student Succeeds Act (ESSA). A meta-analysis was then performed using a random-effects model to provide a statistical synthesis of multiple studies, thereby quantifying an overall effect size, increasing statistical power, resolving discrepancies between individual studies, and drawing more generalizable results. Additionally, we sought to identify factors that moderate the program's impact on science proficiency and broader community outcomes. Each school provided science assessment performance as measured by the Florida State Science Assessment (FSSA for Grades 5 and 8) or the NWEA MAP Science Assessment (4th through 8th grades).

Across all studies reporting Science assessments, the average standardized mean difference was Hedges’ g = 0.68 [95% CI: 0.54, 0.82], Z = 9.51, p < .001. The random-effects model found significant heterogeneity (Q (7) = 260.93, p < .001; I² = 96%), prompting moderator analyses to explore differential impacts across student groups and program fidelity conditions. An analysis to compare the mean effect (Hedge’s g) from schools using the outcome measure of the NWEA Map Science assessment or using the outcome of the Florida State Science Assessment (FSSA) showed average effect sizes g=0.58 [CI% 0.44, 0.73] and g=0.76 [CI% 0.50, 1.03], respectively. The test of the nullhypothesis that these effects were from the same group of effects (Z = 9.63, p < .001) was rejected.

Schools using the Florida State Science Assessment showed a Hedges’ g = 0.76 [95% CI: 0.50, 1.03], Z = 5.7, p < .001. This represents an estimated 28% gain over the control science instruction conditions, can be benchmarked as two-plus years of typical academic learning achieved within a single school year, and resulted in significantly higher passing rates on the FSSA. These effects are statistically rare within the field of education, exceeding the 99th percentile of program impacts reported in Kraft’s (2020) national database of over 700 intervention studies, where the median effect is 0.1 standard deviations.

Further moderator testing revealed robust and positive effects for student subgroups: Title I (g = 0.73), students with disabilities (g = 0.52), and English language learners (g = 0.42). Notably, mastery-based engagement, defined as completing learning activities with an accuracy of≥80%, displayed a dose-response relationship, with low, moderate, and high mastery associated with effect sizes (g) of 0.56, 0.89, and 1.40, respectively. These findings underscore that the quality of engagement, rather than time-on-task alone, predicts the largest achievement gains on independent normed assessments. Together, these results indicate that Penda Learning is both practical and equitable, supporting student progress across demographic and need-based subgroups.

The policy relevance of these findings lies in their dual implications for educational equity and economic vitality. Penda Learning’s effect size surpasses both the What Works Clearinghouse benchmark for a “substantively important” result (≥0.25 sd) and the average gains achieved by most digital learning interventions (Kraft, 2020; 2023, and Bloom et al, 2008). The consistent subgroup effects across ELL, SWD, and Title I populations provide actionable evidence that technology-based mastery learning can help close achievement gaps that have persisted for decades. The magnitude of benefit for Florida schools, particularly the 20–25 percentage-point proficiency improvement observed in FSSA results, suggests that scalable, standards-aligned digital platforms can serve as powerful tools for advancing state ESSA goals and improving districtaccountability metrics.

Collectively, the evidence presented here positions Penda Learning as a promising and scalable science intervention that meets ESSA evidence standards, delivering both substantial educational value and measurable community benefits.