Poster #37 - Narrowing the math achievement gaps with personalized active learning opportunities

• In Event: 2-163 - Poster Session 08
  In Poster Session: PS 08 Section - Cognitive Processes

Integrative Statement

Racial and economic inequality in student achievement is old news, having been documented since the 1960s (Coleman et al., 1966). Much research has been dedicated to understanding the causes and consequences of this gap (Carnervale, 1999), and different ways it can be addressed (Tekian & Hruska, 2004). Importantly, the fact that the achievement gap is a multifaceted issue that involves school, societal, community and home factors, increases the challenges to address and reduce it. In this work, we describe data from a study investigating how an adaptive computer-based tutor improves students grades, especially among minority students, and how adaptive and active learning classroom environments might help reduce the achievement gap.

Computer-based tutors have been shown before to substantially accelerate student learning in mathematics (Koedinger et al., 1997). In fact, a large-scale randomized controlled trial demonstrated that the use of computerized math tutors doubled the rate of math learning in middle school (Pane et al., 2014). However, whether computer tutors have similar effects for minority students and whether their use can contribute to reducing the achievement gap is an open question. It is possible, for instance, that minority or underrepresented students might use the technology less, or interact with the tutor differently and thus not benefit from its use as much as their peers.

In this study we analyze data from over 300 middle school students (6th, 7th, and 8th grade) across four academic years. Students in this sample were enrolled in different schools (0-50% non-white) of a large district in the US. The students in this sample used a cognitive tutor as part of their normal math instruction. The tutor presents students with a series of problems targeting a specific knowledge component and tracks students’ performance on these problems to decide when the skill has been mastered using an underlying cognitive model. Students are provided with extensive feedback and hints when solving the problems. All students completed state-level math proficiency exams.

We found four main results; First, by comparing math performance in the state tests between students who used the computerized tutor and state averages, we found that students who used the computerized tutor performed well above the state average and that this performance grew with years of implementation of the tutor (Figure 1). Second, we found that the use of the computerized tutor improved math test performance among all groups of students. Third, we found that this effect was true for both high minority and low minority schools. Fourth, we found evidence that improvements in state math score is better predicted by opportunity to learn in the tutor than the students’ background. Students who mastered more skills in the tutor and completed more problems showed higher state scores than those who spent less time in the tutor (Figure 2).

Consistent with previous research using other approaches (Haak et al., 2011), these results suggest that interactive, active engagement with adaptive tutors that provide extensive hints and feedback can promote math achievement among minority students and contribute to reducing the achievement gap.

Authors

• Paulo Carvalho, Carnegie Mellon University
  Presenting Author

• Andrés Gómez-Colunga, Yale University
  Non-Presenting Author

• Steve Ritter, Carnegie Learning, Inc
  Non-Presenting Author

• Ken Koedinger, Carnegie Mellon University
  Non-Presenting Author