Abunawas, Mahmoud (2014-12). A Meta-Analytic Investigation of the Predictive Validity of the Test of English as a Foreign Language (TOEFL) Scores on GPA. Doctoral Dissertation. Thesis uri icon

abstract

  • Research conducted on the relationship between international students' TOEFL scores and academic performance has produced contradictory results. To examine this relationship, a meta-analysis was conducted on extant studies centered on international students' TOEFL scores and academic performance as measured through GPA. The meta-analysis included 47 independent effect size values generated from 40 studies. The results of this meta-analysis yielded a statistically significant and positive relationship between international students' TOEFL scores and their GPA. As the variation across effect size values was substantial, a moderator analysis was conducted to detect potential variables contributing to this variation. The moderator variables examined in this meta-analysis included: (a) publication status , (b) graduate level, (c) test version, and (d) research setting. Of these four moderators, only the research setting was found to be a potential moderator. Studies conducted outside the U.S. were found to have higher effect size values than those in the U.S. Implications and recommendations were also discussed.
  • Research conducted on the relationship between international students' TOEFL scores and academic performance has produced contradictory results. To examine this relationship, a meta-analysis was conducted on extant studies centered on international students' TOEFL scores and academic performance as measured through GPA. The meta-analysis included 47 independent effect size values generated from 40 studies. The results of this meta-analysis yielded a statistically significant and positive relationship between international students' TOEFL scores and their GPA.
    As the variation across effect size values was substantial, a moderator analysis was conducted to detect potential variables contributing to this variation. The moderator variables examined in this meta-analysis included: (a) publication status , (b) graduate level, (c) test version, and (d) research setting. Of these four moderators, only the research setting was found to be a potential moderator. Studies conducted outside the U.S. were found to have higher effect size values than those in the U.S. Implications and recommendations were also discussed.

publication date

  • December 2014