Poster #123 - Sex Differences in Growth Curve Model of Early Math Ability

• In Event: 2-163 - Poster Session 08
  In Poster Session: PS 08 Section - Methods, History, Theory

Integrative Statement

Sex differences in learning math have been studied for decades. In 1990, a meta-analysis study reported that differences slightly favoring females in computation in elementary school and middle school, but favoring in males in problem solving in high school and college (Hyde, Fennema, & Lamon). In 2010, a meta-analysis study by Lindberg et al. reported no sex difference in math performance among and in 2015, Reilly et al. reported small but relatively stable differences in math skills favoring males among high school students. While these studies have targeted school-aged students from elementary school to college, little is known about early sex differences in math skills nor whether early sex differences impact developmental changes in math skills among preschool and kindergarten children.
The development of young children’s math skills has been examined in some studies using longitudinal data. Many analyses are based on traditional growth linear modeling and, more recently, nonlinear growth model. Ryoo et al. (2017) estimated changes in math ability score trajectories of preschool children beginning at three years of age over a period of 3 years and at four measurement points. These change trajectories were found to characterize three groups (“fast learners,” “slow but steady learners,” and “high-performing learners”). The Ryoo et al. study, however, did not examine sex differences in performance. The purpose of the current project is to extend the Ryoo, et al. (2017) study by examining the association between sex and the nature of change in math ability. Using the same data set as Ryoo, et al., growth curve analysis (GCA; Bryk & Raudenbush, 1987) techniques within a multilevel modeling framework were implemented to examine whether the rate of change in the math ability score over time was different between males and females.
Multilevel modeling result was reported in Table 1, and the model graph is shown in Figure 1. Adding gender to level 2 in this multilevel model was found to significantly improve the model fit compared with the quadratic change model, χ2 (3) = 9.99, p=.018. Across the participants, there was a random variability in the degree of curvilinear change in math ability scores over time. On average, males’ math ability score at Time 1 (when Time=0) was 83.03, and there was no gender difference in math ability score at Time 1, t=-1.580, p=.115. However, on average, children experienced deceleration (negative curvilinear change) in math ability scores over time, and this deceleration was greater for females, coeff. = -.98, t= -2.29, p=.023, indicating that males scores increased in a faster rate than females.
These findings provided a positive perspective of learning mathematics by demonstrating that there are various change trajectories among young children. Males are fast learners while females are slow but steady learners; however, there is little difference in outcomes. Future studies are encouraged to employ more factors, such as students’ self-efficacy, gender stereotype in learning math, and so forth, to further examine the associations between those factors and the variability of the trajectories of learning math.

Authors

• Keting Chen, University of Nebraska-Lincoln
  Presenting Author

• Victoria J Molfese, University of Nebraska-Lincoln
  Non-Presenting Author