Effects of homework motivation and worry anxiety on homework achievement in mathematics and English

Abstract

Direct and mediating effects of homework worry anxiety on homework effort and homework achievement and the differences in the structural relations among homework motivation constructs and homework achievement across mathematics and English homework were examined in 268 tenth graders in China. Homework motivation included task value, homework self-efficacy, homework worry anxiety, and motivation application. Homework accomplishments were rated by mathematics and English teachers. Homework value had positive effects on homework effort and worry anxiety in both subjects. Homework self-efficacy had positive and negative effects on homework effort and worry, respectively. Homework worry mediated the relation of homework value to effort and to achievement; the relation was more prominent in mathematics than in English homework. The mediating effects of worry anxiety in the relation of self-efficacy to homework effort and to achievement were significant only in mathematics. The domain specificity and direct and mediating effects were discussed in the cultural and educational context.

Key words:

• homework
• motivation
• worry anxiety
• task value
• self-efficacy
• persistence

Notes on contributors

Eunsook Hong is professor emerita of Educational Psychology at the University of Nevada, Las Vegas. Her areas of research interest include self-regulated learning, metacognition, motivation, creativity, and homework. She is an associate editor of the American Educational Research Journal and a member of the editorial board of several journals. Books published include Homework: Motivation and Learning Preferences and Preventing Talent Loss.

Elsa Mason is a doctoral student of Educational Psychology at the University of Nevada, Las Vegas. Her research interests include expectancy-value interventions, metacognition, and self-regulation. She is also a psychology professor at the College of Southern Nevada and teaches courses in general psychology, personal adjustment, human relations, and research methods. She has held positions in student affairs and academic affairs and led a team of specialists who were charged with assisting at-risk students persist in college and graduate in a timely manner.

Yun Peng is a doctoral student of Educational Psychology at the University of Nevada, Las Vegas. Her research focuses on three primary areas: study strategies, creative thinking, and homework. Research method is another area of concentration, including qualitative, quantitative, and mixed research methods, with advanced statistics such as multivariate analysis and structural equation modelling.

Nancy Lee received her PhD degree in Learning and Technology from the University of Nevada, Las Vegas with a dissertation on the effects of computer programming learning strategies. Her research interests include various learning strategies with the focus on self-explanation. She teaches Computer Science and Web Design and Development at Advanced Technologies Academy while providing training to the school district's Career and Technical Education teachers and to the elementary school teachers for the implementation of computer science education through the Code.org curriculum.

Notes

1. Model adequacy was evaluated by a number of fit indexes reported by the EQS program (Bentler, 2008). A value of CFI and NNFI greater than .95 indicates an acceptable fit to the data (Hu & Bentler, 1999). A value of an SRMR less than .08 is considered a good fit. A model with an RMSEA of smaller than .08 indicates a good fit, .08 to .10 a mediocre fit, and larger than .10 a poor fit (MacCallum, Brown, & Sugawara, 1996). Although χ² values are routinely reported, due to its sensitivity to sample size (e.g., if same size is larger than about 200, χ² is almost always statistically significant, indicating poor fit), readers are to interpret χ² findings with caution. The LM test evaluates whether some of the fixed parameters in the model could be freed, while the Wald test is used to evaluate whether some of the free parameters in the model could be restricted.