How Noncognitive Constructs Perform Across Grades and Subjects and Vary by Student Subgroup

• In Event: Noncognitive Constructs and Academic Outcomes: Using NAEP Data to Understand Differences in Achievement

Abstract

There is increasing interest in noncognitive constructs, as they are suggested to hold an important relationship with academic achievement and other student outcomes (Farrington et al., 2012; Heckman & Kautz, 2012; Richardson et al., 2012). Previous research finds constructs such as academic motivation, interest, and self-efficacy to be related to mathematics and reading performance (Retelsdorf et al., 2011; Richardson et al., 2012; Bohrnstedt et al., 2020; Zhang et al., 2021).

Due to a lack of representative samples in most previous research, there is a need to provide more empirical evidence on the relationships of these constructs to academic achievement using nationally representative data. This study used 2019 NAEP data to provide an overview of these constructs, whether and how they differ by student group within each grade/subject, and whether their relationship to NAEP achievement is similar or different across subjects and grades.

Data for this study is derived from the public-school sample of the NAEP 2019 grade 4 and 8 reading and mathematics assessments and NAEP student questionnaire (see Tables 1 and 2). This analysis includes seven noncognitive constructs (see Appendix A for indices and items). The current study used descriptive and correlation analysis to examine the overall means and standard deviations for the noncognitive construct indices and subgroup comparisons by gender, race/ethnicity, and NSLP status for each grade/subject. We computed standardized differences for the subgroup differences, and effect sizes equal to or bigger than 0.2 are highlighted in the results.

Results show that confidence in reading/mathematics knowledge/skills was moderately correlated with performance across both grades and subjects. In contrast, academic self-discipline was weakly correlated with performance, and the motivation to meet performance goals was weakly or not correlated with performance across grades and subjects. The correlation magnitudes were similar across grades and subjects (see Table 3).

Examining subgroup differences across both grades and subjects, female students had higher scores on the self-discipline index than male students. In reading, for both grades 4 and 8, female students also scored higher on reading confidence, interest, and motivation to meet mastery goals than male students, and in grade 4, female students also had higher scores on the persistence in learning index.

Across both grades and subjects, White students consistently reported higher scores on confidence in reading/mathematics knowledge and skills than Hispanic students. Black students reported higher scores than White students on various indices across subjects, such as the performance goals index in grade 4 and mastery goals index in grade 8. In grade 8, Asian students scored higher on interest and mastery goals than White students. Finally, non-NSLP students scored higher on self-discipline and confidence across both subjects and grades than NSLP students (see Tables 4 – 23 for all subgroup comparisons).

This study is a significant step in understanding the relationship between student noncognitive constructs and achievement and important patterns in the differences across indices for student subgroups. The following presentation expands upon this work by addressing the interaction effects of the noncognitive constructs in predicting performance with student groups.

 

Authors

• Alison Filbey, American Institutes for Research
  Presenting Author

• Grace Ji Cheng, American Institutes for Research
  Non-Presenting Author

• B. Jasmine Park, American Institutes for Research
  Non-Presenting Author

• Jennifer Mazzella
  Non-Presenting Author