skip to main content skip to footer

Measuring Self-Concept and Relating It to Academic Achievement: Statistical Analyses of the Marsh Self-Description Questionnaire

Wenglinsky, Harold
Publication Year:
Report Number:
ETS Research Report
Document Type:
Page Count:
Subject/Key Words:
Academic Achievement, Intervention, Marsh Self-Description Questionnaire, Performance Factors, Self Concept


Educational intervention programs that seek to raise student self-concept as a lever for raising student achievement are hampered both by limitations in measurement tools and a lack of knowledge about the relationship between self-concept and student achievement. Advances in self-concept research indicate that student self-concept includes many domains corresponding to the various endeavors of adolescents, from relations with peers to physical abilities. The most state-of-the-art instrument for adolescents, which takes the multiple domains of self- concept into account, is the Marsh Self-Description Questionnaire II (SDQ II). Yet the instrument is lengthy, containing 102 questions including some that are regarded by teachers as insensitive to the feelings of their students, creating problems in administering the instrument. Further, little research has been done on the relationship between specific domains of self- concept and academic achievement; most researchers tend to claim either that all domains of self-concept affect academic achievement, or that self-concept is entirely a by-product of previous student performance, and therefore does not affect it at all. This study addresses these two issues. First, it applies confirmatory factor analysis to a database constructed by the author from an administration of the 102-item SDQ II to a sample of 1,130 students participating in Outward Bound, USA programs, finding that the instrument can be reduced to thirty items with little loss of reliability. Second, it applies regression analysis to the National Educational Longitudinal Study (NELS), which contains items from the SDQ II, to measure the relationship between each self-concept domain and achievement in reading and mathematics. This analysis finds that the effect of self-concept on achievement is domain specific; academic domains of self-concept affect academic achievement, while nonacademic domains do not. The study concludes with some suggestions for how intervention programs such as Outward Bound, USA can measure self-concept and use increases in specific types of self-concept as a lever for raising student achievement.

Read More