DEDICATIONThis dissertation is dedicated to my husband, Craig, for his unwavering belief in me and complete support of all my professional endeavors.iv ACKNOWLEDGEMENTS I would like to thank my dissertation advisor, Dr. Tom Tretter, for his patient guidance, enthusiastic encouragement, and wise counsel. I would also like to thank my dissertation committee; I have learned invaluable lessons from each one of you that I will do my best to pass on to my future students. I thank my children, Susannah, Kelly, and Alex, who have inspired me to be a life-long learner. Their enduring patience with Mom's incessant planning, writing, and studying over the last four years has not gone Increasing retention of students in science, technology, engineering, or mathematics (STEM) programs of study is a priority for many colleges and universities.This study examines an undergraduate teaching assistant (UTA) program implemented in a general chemistry course for STEM majors to provide peer learning assistance to entrylevel students. This study measured the content knowledge growth of UTAs compared to traditional graduate teaching assistants (GTAs) over the semester, and described the development of peer learning assistance skills of the UTAs as an outcome of semesterlong training and support from both science education and STEM faculty. Impact of the UTA program on final exam grades, persistence of students to enroll in the next chemistry course required by their intended major, and STEM identity of students were estimated. The study sample comprised 284 students in 14 general chemistry recitation sections led by six UTAs and 310 students in 15 general chemistry recitation sections led by three traditional GTAs for comparison.Results suggested that both UTAs and GTAs made significant learning gains in general chemistry content knowledge, and there was no significant difference in content vi knowledge between UTA and GTA groups. Student evaluations, researcher observations, and chemistry faculty comments confirm UTAs were using the learning strategies discussed in the semester-long training program. UTA-led students rated their TAs significantly higher in teaching quality and student care and encouragement, which correlated with stronger STEM recognition by those students.The results of hierarchical linear model (HLM) analysis showed little variance in final exam grades explained by section-level variables; most variance was explained by student-level variables: mathematics ACT score, college GPA, and intention to enroll in the next general chemistry course. Students having higher college GPAs were helped more by having a UTA. Results from logistic regression of persistence outcome variable showed that students are three times more likely to persist to CHEM 202 if they had a UTA in CHEM 201. Other positive predictors of retention included having strong college grades, and having strong ACT math scores. Coupled with HLM analysis result that UTAs were more effective at helping students with higher college GPAs achieve higher grades, t...