A Randomized Controlled Trial (RCT) of...
Bernie’s Book Bank’s Free Book Distribution Program for Low-Income K-6 Students
Reviewed
PNAS (April 2026) published an RCT of Bernie’s Book Bank’s free book distribution program for K-6 students in high-poverty schools in Milwaukee. The study reports positive impacts on reading achievement, but these findings are based on analysis of a subgroup that was defined post hoc (i.e., wasn’t prespecified). Meanwhile, results of the study’s prespecified primary analysis aren’t reported, preventing reliable conclusions about program effectiveness.
For more than a century, studies have shown that children who grow up in homes with more books achieve higher levels of academic success, yet it remains unclear whether books themselves improve learning or simply reflect broader socioeconomic advantages. Skill-development theory holds that greater access to books directly improves literacy through increased print exposure and reading practice, whereas the cultural capital account suggests books are indicators of broader family resources, including parental education, academic norms, and enrichment opportunities, that promote achievement independently of the books themselves. To provide causal evidence, we conducted a school-level randomized controlled trial of a program that builds children’s home libraries. In 2018, we randomly assigned 60 high-poverty public elementary schools to treatment or control groups. Students in 30 treatment schools received four book distributions over 5 y, averaging about seven books per distribution, and prominently including high-interest and culturally relevant titles; students in 30 control schools received none. Tracking students from 2018–19 through 2022–23, we find a statistically significant intention-to-treat impact on reading achievement of d = 0.100 and a larger d = 0.207 advantage for those completing the full 5-y program. These impacts correspond to approximately 25 to 32% and 52 to 65% of a typical year’s learning, respectively. The largest benefits are concentrated among students who received books across all distributions, indicating that cumulative exposure drives the strongest impacts. These findings provide evidence supporting skill-development theory and highlight a scalable strategy for improving literacy outcomes in high-poverty urban schools.
For more than a century, studies have shown that children who grow up in homes with more books achieve higher levels of academic success, yet it remains unclear whether books themselves improve learning or simply reflect broader socioeconomic advantages. Skill-development theory holds that greater access to books directly improves literacy through increased print exposure and reading practice, whereas the cultural capital account suggests books are indicators of broader family resources, including parental education, academic norms, and enrichment opportunities, that promote achievement independently of the books themselves. To provide causal evidence, we conducted a school-level randomized controlled trial of a program that builds children’s home libraries. In 2018, we randomly assigned 60 high-poverty public elementary schools to treatment or control groups. Students in 30 treatment schools received four book distributions over 5 y, averaging about seven books per distribution, and prominently including high-interest and culturally relevant titles; students in 30 control schools received none.
In an exploratory analysis of the subgroup of students in first and second grade at study entry, tracked
Tracking students from 2018–19 through 2022–23, we find a statistically significant intention-to-treat impact on reading achievement of d = 0.100 and a larger d = 0.207 advantage for those completing the full 5-y program. These impacts correspond to approximately 25 to 32% and 52 to 65% of a typical year’s learning, respectively. The largest benefits are concentrated among students who received books across all distributions, indicating that cumulative exposure drives the strongest impacts. These findings
warrant examination in future research, as they may
provide evidence supporting skill-development theory and highlight a scalable strategy for improving literacy outcomes in high-poverty urban schools.
No-Spin’s Study Overview
An RCT of Bernie’s Book Bank’s free book distribution program for K-6 students in high-poverty Milwaukee schools doesn’t report findings of its prespecified primary analysis, preventing reliable conclusions about program effectiveness. Positive findings from an exploratory subgroup analysis may warrant examination in future research.
Program:
- Bernie’s Book Bank’s program provided free distributions of “new and gently used, high-interest and culturally relevant children’s books” to all K-6 students in participating schools.
- As implemented in the study, students received four book distributions over five years, averaging approximately seven books per distribution. The COVID-19 pandemic limited implementation in two of the five study years.
Study Design:
- The study randomly assigned 60 high-poverty Milwaukee elementary to either a treatment group (N=30) in which all K-6 students received free book distributions, or a control group (n=30) that did not receive book distributions.
- Per study’s preregistration, “The primary intention-to-treat (ITT) research question we will address is: Does school-level assignment to the Bernie’s Book Bank program impact the reading achievement outcomes of the schools’ first-through-fifth grade students?” See the preregistered statistical analysis and related documents [1, 2].
Findings:
- The study doesn’t report results of its prespecified primary analysis focused on students in grades 1-5, preventing reliable conclusions about program effectiveness.
- The study instead reports positive results of a post-hoc (not prespecified) analysis, focused on the subgroup of students in grades 1-2 at study entry and making other material departures from the prespecified plan – including changing the school-level control variables (“covariates”) and method for handling missing data.
- Such post-hoc findings are considered exploratory under established standards (IES, FDA), generating hypotheses that may be useful to examine in future research.
Comment:
- Disclosure: Arnold Ventures, which funds No-Spin Evidence Review, funded this RCT of Bernie’s Book Bank.
Click or tap a highlight to see No-Spin’s comment