A new groundbreaking data analysis found that using Speak Agent Math+Language℠ leads to significant gains in math scores for middle school students at Prince George’s County Public Schools. The research also found a significant effect on English language proficiency.
Summary of Findings
PGCPS implemented Speak Agent in 1,600 middle grades math classrooms during SY 2021-2022. Nonprofit research organization Leanlab Education evaluated the impact of program usage on both math and English language proficiency outcomes:
- Cross-sectional analysis of 46,929 quarterly Math Benchmark test scores found that, on average, completing 10 Speak Agent activities corresponded with an increase of 10% on the Math Benchmark score compared to students who completed zero activities in Speak Agent.
- Time-series analysis of the same 46,929 Math Benchmark test scores over three academic quarters found that, on average, an increase in the number of Speak Agent activities a student completes will increase the math test score by 0.6% per activity, from one quarter to the next quarter.
- Cross-sectional analysis of 5,361 WIDA ACCESS for ELLs test scores found that, on average, completing 10 Speak Agent activities corresponds with an increase of 17 points on the Overall Scale Score.
What Makes This Research Newsworthy?
Let's unpack what is "groundbreaking" about this research. First, this is an exceptionally large sample size for an edtech program, with more than 24,000 students in 1,600 classrooms. As a result, the key findings are significant at an extraordinarily high 99.9% confidence level (p<0.001).
The effect sizes are also large. Completion of just 10 Speak Agent learning activities within an academic quarter corresponds to 10.3% higher math scores. This is easily achievable within our recommended dosage of one hour per week. Critically, outcomes were measured on independent assessments developed by third-parties (Pearson and WIDA) that have no affiliation with Speak Agent, Inc.
Furthermore, unlike most research studies, this analysis observed usage in real-world classrooms where teachers and students had a choice of whether or not to use Speak Agent — and how frequently. The student population is highly diverse, and all groups benefited in math, ELP, or both:-
- Black: 55.4%
- Latino: 37.3%
- Multilingual: 16.0%
- Title I: 38.3%
- FARM (Free and Reduced-Price Meals): 67.6%
Last but not least, this research found a significant effect on math and English language proficiency using one single program. This research highlights a solution for school districts with stretched resources that can only implement a limited number of programs in a given school year.
Is Speak Agent Done with Research?
As a team of educators, nothing is more important to us than understanding why our products improve learning outcomes and how we can take that to the next level. We will never stop supporting independent, third-party research!
Read Also
- Digital Promise: Integrating Research-based Strategies into the Design of Educational Technology to Support All Learners
- ClassTech Tips Blog: How to Learn Math and Science with Language Support
- eSchool News: How a Major MD District Boosted Middle Math Achievement
- The Evidence Basis: A Summary of Impact Studies
- Speak Agent Launches Formative Assessment Tool
- Can Edtech Help with the Teacher Time Crunch?
- Speaking in Crisis
- Speak Agent Awarded Learner Variability Product Certification
- Speak Agent Awarded Research Based Design Product Certification
Written by Ben Grimley
Ben is CEO and Co-Founder of Speak Agent, Inc.
