Dissertation Defense Announcement
College of Education announces the Final Dissertation Defense of
for the Degree of Doctor of Education
April 5, 2019 at 10:00 AM in Ball Hall, Room 320
Advisor: Andrew Tawfik
Computational Thinking Self-Efficacy in High School Latin Language Learning
ABSTRACT: Research suggests that computational thinking is a necessary skill exercised in STEM courses, non-STEM fields, and in everyday life. However, very little research has investigated the potential transfer of computational thinking self-efficacy available through classical Latin courses. This causal comparative study contrasted the computational thinking self-efficacy of computer science students with no exposure to Latin to computer science students with exposure to Latin at a Memphis all-boy high school. The independent variables were Latin language learning experience, i.e., up to 6 total of Latin language learning (n = 33), versus 0 years of Latin language learning experience (n = 20). Additional data on the number of years enrolled in other foreign languages was collected. The dependent variable was mean scores of items found on a computational thinking and problem solving self-efficacy scale. This instrument uses a Likert scale to measure students' self-efficacy in nine computational thinking components including algorithmic thinking; abstraction; problem decomposition; data collection, representation, and analysis; parallelization; control flow; incremental and iterative; testing and debugging; and questioning. To test the null hypothesis that having a Latin language learning yields no significant influence on computer science students' self-efficacy in computational thinking and problem solving, a multivariate analysis of variance (MANOVA) test was utilized for this causal-comparative study. To test the null hypotheses that having a Latin language learning yields no significant influence on computer science students' abstraction, problem decomposition, data, parallelization, control flow, incremental and iterative, testing and debugging, and questioning skills self-efficacy, a separate ANOVA test were run for each computational thinking skill component. The data did not meet of the necessary assumptions for a MANOVA test. The sample size for the Latin group was of no concern at n = 33. The means from the descriptive statistics show that the non-Latin group outscored the Latin group in most of the computational thinking skills. Pillai's trace statistic from the MANOVA test showed no statistical significance in the computational thinking and problem solving scale. The individual results from the ANOVA tests showed no statistical significance for any of the nine subscales.