Unwarranted Uncertainty: Gender Patterns in Early Mathematical Confidence, Interest, and Achievement

• In Event: Gender and Mathematics: An Interdisciplinary Symposium Examining Classroom Influences, Achievement, and Affect

Abstract

Given that gender differences in STEM-related career choices are substantial and larger than gaps in achievement, there are clearly factors other than performance that contribute to career disparities. One potential factor implicated in much research is student affect (e.g., Forgasz, Leder, & Kloosterman, 2004; Hyde, et al, 1990; Leder & Forgasz, 2002; Meece, Glienke & Burg, 2006).

Although affect is defined in various ways, this study focuses specifically on students’ confidence and interest in math, examining the following questions:
1. Are there differences in girls’ and boys’ mathematical confidence, interest and achievement?
2. Does early mathematical confidence or interest predict later achievement gains? Does prior achievement predict later confidence and interest? Are patterns similar for boys and girls?

This study used ECLS-K on roughly 10,000 U.S. students followed from K-8th grades. At grades 3 and 5, students indicated their agreement (1-4 scale) with 4 statements relating to their math confidence (e.g., I am good at math.) and 4 others regarding their interest in math (e.g., I like math). Slightly different questions were asked at grade 8 (resulting cautions will be discussed in the full paper).

We examined gender disparities in confidence, interest, and achievement using t-tests. Additionally, we constructed three series of regression models with 5th-grade achievement, confidence, and interest serving as the outcomes, and 3rd grade demographics, confidence, interest, and prior achievement serving as predictors. We then ran three similar models examining 5th-8th grades. Interaction terms were included to examine whether patterns were similar for boys and girls.

T-test results indicated that the gender gap in math confidence was substantially larger than disparities in both student interest and actual math achievement. For example, the effect sizes of boys’ advantage in math confidence, interest and achievement were .33, .14, and .19, respectively at grade 3, and .29, .07, and .23 at grade 5.

Regression models revealed that math confidence—but not interest—predicted growth in achievement during middle (but not elementary) school for both girls and boys. Additionally, math confidence at 3rd and 5th grades predicted interest in math at 5th and 8th grades, respectively. Over 70% of the variation in achievement was “explained” by prior achievement, attitudes, and demographics in the regression models; in contrast, only 20% of the variation in student interest and confidence was explained. Hence, students’ math confidence and interest appear to be more dependent upon their year-by-year experiences.

In summary, U.S. girls feel substantially less mathematically competent than do boys, with disparities in confidence greater than actual achievement differences. Given that early math confidence was found to be a significant predictor of middle school gains in math achievement and interest, these patterns raise concerns about ways in which young girls perceive their own math abilities and the impact this might have on future course taking and career decisions. This study suggests that, instead of focusing on girls’ achievement and interest, gender-focused interventions might be most fruitful if they target girls’ perceptions of their math abilities and include long-term supports to help maintain girls’ confidence in math.

Authors

• Sarah Theule Lubienski, University of Illinois at Urbana-Champaign

• Colleen M. Ganley, University of Illinois at Urbana-Champaign

• Corinna C. Crane, ECRA Group