Understanding How Ability Stereotypes Predict Girls' Math Self-Concepts: A Longitudinal Investigation

• In Event: Understanding Students' Math Self-Concept: Its Relation With Achievement, Development, and Educational Implications

Abstract

Objectives. A child’s math self-concept is a powerful predictor of their math achievement and choices (Marsh et al., 2005). Math self-concept typically declines between elementary and high school (Cvencek et al., 2021), and this decline is stronger for girls than boys (Sáinz & Eccles, 2011). Cross-sectional studies provide a possible explanation for the larger decline in girls’ self-concept. The societal stereotypes that “boys are better in math” may impede girls from developing beliefs in their math competence during early adolescence (Cvencek et al., 2021; Sáinz & Eccles, 2012). However, it is important to longitudinally examine the link between girls’ gender-ability stereotypes and math self-concept to examine how these stereotypes influence the development of math self-concepts over time (Parker et al., 2015). Our primary aim was to examine how math gender-ability stereotypes influenced girls’ math self-concepts across multiple school levels (elementary and middle/high school) from a longitudinal perspective.

Theoretical Framework. As a theoretical framework, we use Marsh and colleagues’ (2005) reciprocal effects model of academic self-concept, which examines how self-concepts and achievement influence one another over time in the context of societal stereotypes.

Method. The current sample included N=422 students (7–14 years old at Time 1, Mean=11.49, SD=1.98; 48% girls) who completed an online survey three times over a 12-month period. Math self-concepts were measured with two items (αs=.87–.89, e.g., “I am good at math activities,” 1-6 scale). Stereotypes were measured with a difference score, “how good are most boys at math” minus “how good are most girls,” 1-6 scale; difference score range -5 to 5. We conducted a two-level linear growth model using Stata15.0 to analyze a data structure where students (level-2) were repeatedly measured across three time-points (level-1). Model testing proceeded in three phases: the unconditional model, the unconditional growth model, and the conditional model (see Supplementary Material).

Results. Results are shown in Table 1. First, average change in math self-concept was not significant across time. Second, the significant three-way interaction of gender×school-level×gender-ability stereotypes across time indicated that gender and school-level significantly moderated the influence of gender-ability stereotypes on the trajectory of students’ math self-concepts. Breaking down this interaction, we examined the simple effect of time on math self-concept among girls who endorsed ability stereotypes favoring boys or favoring girls at different school levels (elementary- and middle/high-school). The only significant decrease in math self-concepts over time was found among elementary-school girls who endorsed stereotypes favoring boys (Figure 1).

Significance. This study examined the influence of gender-ability stereotypes on math self-concept among students in elementary and secondary schools across a 12-month period. We found a significant association between girls’ math gender-ability stereotypes and self-concept longitudinally. Elementary-school girls who endorsed ability stereotypes that favored boys were more likely to have a decreasing trajectory in math self-concept. These findings demonstrate how gender stereotypes and biases can have a negative impact on young girls’ math self-concept development. Taken together, our study highlights the need to protect girls’ self-concepts from negative gender stereotypes about their abilities and promote their motivation in math.

Authors

• Daijiazi Tang, University of Houston

• Sapna Cheryan, University of Washington - Seattle

• Andrew N Meltzoff, University of Washington

• Allison Master, University of Houston