Paper Summary

Direct link:

Longitudinal Links Between Task Values and Math Grades: Considering Different Time Lags Between Measurement Occasions

Sat, April 18, 10:35am to 12:05pm, Virtual Room


Purpose and Theoretical Framework
In line with expectancy-value theory (EVT; Eccles et al., 1983), longitudinal studies have shown that students with comparatively higher domain-specific task values subsequently achieve better in this domain, even if prior achievement is controlled (e.g., Steinmayr & Spinath, 2009). However, recent research suggests that the chosen time lag is very important when testing relations between two constructs over time (Dormann & Griffin, 2015). As values refer to the current emotional and cognitive evaluation of a given subject (see Eccles, 2005), considering optimal time lags might be particularly critical when the learning contexts change. The present study investigates whether the relation between values and grades changes when considering different time lags and changing learning conditions.

A sample of 476 German students were followed for 3 years from the beginning of Grade 11 (M = 16.43 years, SD = 0.55) to the end of Grade 13. Teachers changed after 11th grade. Students reported their task values (intrinsic, attainment, and utility values) in math at the beginning of Grade 11 (t11.1) and in the second term of Grades 11, 12 and 13 (t11.2, t12.2, t13.2). Schools provided all report card grades in math for Grades 10 through 13 (i.e., 7 report cards). Covariates were assessed at t11.1 (students’ cognitive abilities, socio-economic status [SES], and gender) or t12.1 (basic vs. advanced math course). First, we tested a latent cross-lagged model (CLM) with all task values assessed at four measurement occasions at an interval of 6 month and 1 year, respectively, and math grades assessed at 7 measurement occasions at an interval of six months (“short lag model”). Second, we estimated a CLM with a testing interval of 2.5 years (“long lag model”) with all task values and grades assessed at t11.1 and t13.2.

In the “short lag model”, math grades significantly predicted changes in task values within Grades 11, 12 and 13 above and beyond all covariates (Figure 1). Effects were small to moderate. Students’ values significantly positively predicted changes in math grades in Grades 11 and 13 and marginally in Grade 12 (p = .051). In the “long lag model”, students’ math grades in Grade 11 significantly predicted changes in task values in Grade 13. By contrast, task values in Grade 11 were not significantly associated with changes in grades from Grades 11 to 13 (Figure 2). Considering the three value components in separate models did not change the pattern of results.

First, findings suggest that task values are relatively malleable even at the end of secondary education, which is a necessary pre-condition for implementing successful value interventions at this educational stage. Second, values seem to be more important for changes in math grades when time lags between measurements are comparatively smaller. Accordingly, short-term assessments of students’ motivational beliefs are necessary to better understand the relations of task values and academic achievement in changing learning contexts.