Predicting Students' Academic Performance in College Using a New Non-cognitive Measure

Download or read book Predicting Students' Academic Performance in College Using a New Non-cognitive Measure written by Sui Huang and published by . This book was released on 2011 with total page 178 pages. Available in PDF, EPUB and Kindle. Book excerpt: Abstract: There have been voluminous studies upon the potential predictive power of students' non-cognitive attributes such as self-confidence, motivation and student's interaction with institutions on their collegiate performance. While much effort have been put towards identifying the utilities of these attributes, little endeavors have been made to design a concise yet precise composite measure of all the collegiate performance-related non-cognitive attributes. Such measure could be beneficial not just to the future research studies about non-cognitive correlates of students' academic performance in college; it could also be more than informative to the admission officials in higher institutions. The initial intent of this study was to develop an instrument to measure relevant non-cognitive drivers of students' academic performance in college. A Meta analysis was conducted to guide the selection of non-cognitive constructs to include in the proposed instrument. Academic self-efficacy and achievement motivation were the two constructs that were shown to be significantly correlated with students' academic performance in college across research studies. Considering the momentum that student engagement is acquiring in evaluating the quality of higher institutions and predicting students' success in college, the author also included student engagement as one of the constructs to measure. Items were created for each of the aforementioned constructs by reviewing, selecting and revising the existing measures. Face and content validity were established through expert reviews. A forty item initial survey were sent out electronically in two mid-west higher institutions. 497 responses from two Ohio higher institutions were collected. Reliability of the proposed instrument was initially established based on Cronbach's alpha. Classical Test Theory and Item Response Theory as well as Exploratory Factor Analysis were conducted in the refinement of the instrument. Confirmatory Factor Analysis was utilized to validate the construct validity of the final survey, which was consist of the 20 selected items from the initial survey. A structural equation exploration was conducted after establishing the reliability and validity of the proposed survey with the intent of studying the added values of non-cognitive attributes on students' cumulative GPA in college after controlling for the influence of their prior success, which was measured by their high school GPA and SAT scores. Two hypothesized models were tested and compared. Results indicated that students' non-cognitive constructs (i.e. Academic self-efficacy, achievement motivation and student engagement) have significant influence on students' cumulative GPA in college while not taking into account their prior academic success. However, when students' prior academic success was included as covariate in the model, most of the aforementioned non-cognitive attributes were no longer significant predictors of their academic performance in college, which in other words could be interpreted that the aforementioned students' non-cognitive constructs are, to some degree, proxies of their prior academic success, which ultimately influence their performance in college.