It's Easy but Unimportant: How Adolescents' Expectancy-Value Motivational Profiles Predict Their Different Reading Outcomes

• In Event: Person-Centered Research: A Methodological Answer to Motivation Theory's Most Critical Questions

Abstract

Background. Expectancy-value theorists posit that students’ motivation to complete tasks is determined jointly by how well they expect to do and how much they value the tasks (Wigfield, Tonks, & Klauda, 2009). Researchers have explored how students’ expectations, values, and their interactions predict achievement and choice (e.g., Trautwein et al, year). However only a few researchers (Guthrie, Coddington, & Wigfield, 2009) have explored how patterns of expectations and values within students predict their reading achievement. This study aimed to (a) find profiles of students’ “affirming” expectancy beliefs and values for reading information texts and “undermining” beliefs and values (perceptions that reading is difficult, devaluing reading (Wigfield, Cambria, & Ho, 2012), and (b) use profile membership to predict three types of achievement outcomes.

Methods and Results. Participants included 876 7th grade students, 71.4% White, 22.7% African American, 46.3% female. Students’ reading self-efficacy, value, devalue, and perceptions of difficulty were assessed twice in a six week period. At the second data collection point we assessed students’ information text comprehension and dedication to reading (defined as students desiring to commit time and effort to reading activities) and collected students’ 4th-quarter reading grades. Ward’s method of agglomerative cluster analysis and k-means analysis were used to identify different clusters of students based on the motivation variables. A four-cluster solution best fit the data. Each cluster had approximately equal membership, and cluster membership explained 50.4% of variance in self-efficacy, 58.6% in valuing, 64.7% in devaluing, and 59.6% in perceived difficulty. Based on the mean unstandardized and standardized values of each variable within each cluster, we named the clusters: high-efficacy-high-value; high-efficacy-high-devalue, moderate-difficulty-moderate-value; and high-difficulty-high-devalue.

One-way ANCOVAs, controlling for ethnicity, gender, and previous performance, were used to assess differences in reading dedication, comprehension and grades based on cluster membership. For dedication all differences between clusters were significant and scores (high to low) were: high-efficacy-high-value (M=3.31, SE=0.04), moderate-difficulty-moderate-value (M=3.07, SE =0.04), high-efficacy-high-devalue (M=2.77, SE=0.05), high-difficulty-high-devalue (M=2.57, SE=0.05). For information text comprehension and grades, high-efficacy-high-value and high-efficacy-high-devalue students scored equally high (ITC: M=17.60, 17.58; SE =0.28, 0.31; grades: M=4.29, 4.25; SE=0.07, 0.08), and significantly higher than students in the other two clusters did (ITC: M=15.27, 15.80; SE=0.30, 0.26; grades: M=3.95, 3.94; SE=0.08, 0.07).

Discussion. Results complement findings from variable-centered studies showing positive relations of expectancies and values to different facets of students’ engagement and performance. They extend them by showing how groups of students with high self-efficacy but who varied in valuing of reading earned similar achievement scores, but had different dedication to reading. Since it is difficult to detect interactions between expectations and values in correlational, variable-centered studies (Trautwein et al., 2012), person-centered approaches like this may be especially fruitful for understanding the predictive relationships of students’ patterns of expectations and values for different reading outcomes. Person centered approaches also may provide insights into how to design effective interventions for groups who have different patterns of expectancies and values. Results provide support for expectancy-value theorists’ proposals that different patterns of expectancies and values might differentially affect outcomes (Wigfield, 1994).

Authors

• Emily Quinn Rosenzweig, University of Maryland - College Park

• Allan L. Wigfield, University of Maryland - College Park