Can Bold Problem Solving and Spatial Skills Explain the Gender Gap in Problem-Solving Performance?

• In Event: “Bold Problem Solving” and Spatial Skills: Two Important Constructs for Gender Equity in Mathematics Education

Abstract

The gender gap in mathematics is evident when children begin formal schooling, persists throughout, and is particularly striking at the top of the distribution, among high-achieving students (Fryer & Levitt, 2010; Cimpian, et al., 2016). Several studies have found that spatial skills help explain gender differences in mathematics (Casey et al., 1995; Halpern et al., 2007; Mix & Cheng, 2012). However, no study has examined the role of “Bold Problem Solving” in explaining gendered patterns in mathematics outcomes.

Given gender differences in spatial skills and bold problem-solving approaches, this study examines:
1) whether there are gender differences in problem-solving performance among high-achieving boys and girls just before high school, and
2) whether spatial skills and/or bold problem-solving approaches account for any of the observed gender difference in problem-solving performance.

This project was part of the study introduced in Paper 2, using data collected from 76 8th graders at a selective 5-year high school. The students completed a problem-solving assessment including five challenging SAT items with known sex differences (Gallagher, 1992), a 3-D Mental Rotation Test (24 items; Peters et al., 1995, based on Vandenberg & Kuse, 1978), and a survey regarding their math beliefs, attitudes, and practices, which included the newly-developed Bold Problem Solving items.

Despite the boys and girls in the sample having similarly high prior math test scores, there were striking gender differences in mental rotation (d=1.18), Bold Problem Solving (d=1.09) and performance on the five challenging SAT items (d=.75). In zero-order correlations, all variables were significantly correlated with one another (see Table 3).

To examine research question 1, we ran a regression analysis in MPlus (Muthen & Muthen, 1998-2012) predicting problem-solving performance from gender. Results showed a large gender difference in problem-solving performance (p=.001) (see top of Figure 1). To test research question 2, we ran a path analysis in MPlus and tested the indirect relations between gender and problem-solving performance through both mental rotation and Bold Problem Solving. In this model, relations between gender and Bold Problem Solving (p<.001) and gender and mental rotation (p<.001) are both statistically significant (see bottom of Figure 1). However, only Bold Problem Solving predicts problem-solving performance (p<.001). Moreover, there is a significant indirect relation between gender and problem-solving performance through Bold Problem Solving (point estimate=0.68, p=.001), but not through mental rotation (point estimate=.27, p=.13). Results hold when we covary out prior SSAT verbal and math scores as well as when using Poisson regression to account for the fact that problem-solving performance is a count variable.

Results suggest that boys’ use of bolder problem-solving strategies helps explain the gender difference in problem-solving performance. In the symposium, we will discuss additional analyses exploring whether Bold Problem Solving mediates the relation between mental rotation and problem-solving performance, and the role of confidence and mental math in these relations.

Authors

• Jennifer Timmer, University of Illinois at Urbana-Champaign

• Colleen M. Ganley, Florida State University

• Sarah Theule Lubienski, Indiana University