abstract: Accurate data analysis and interpretation of results may be influenced by many potential factors. The factors of interest in the current work are the chosen analysis model(s), the presence of missing data, and the type(s) of data collected. If analysis models are used which a) do not accurately capture the structure of relationships in the data such as clustered/hierarchical data, b) do not allow or control for missing values present in the data, or c) do not accurately compensate for different data types such as categorical data, then the assumptions associated with the model have not been met and the results of the analysis may be inaccurate. In the presence of clustered/nested data, hierarchical linear modeling or multilevel modeling (MLM; Raudenbush & Bryk, 2002) has the ability to predict outcomes for each level of analysis and across multiple levels (accounting for relationships between levels) providing a significant advantage over single-level analyses. When multilevel data contain missingness, multilevel multiple imputation (MLMI) techniques may be used to model both the missingness and the clustered nature of the data. With categorical multilevel data with missingness, categorical MLMI must be used. Two such routines for MLMI with continuous and categorical data were explored with missing at random (MAR) data: a formal Bayesian imputation and analysis routine in JAGS (R/JAGS) and a common MLM procedure of imputation via Bayesian estimation in BLImP with frequentist analysis of the multilevel model in Mplus (BLImP/Mplus). Manipulated variables included interclass correlations, number of clusters, and the rate of missingness. Results showed that with continuous data, R/JAGS returned more accurate parameter estimates than BLImP/Mplus for almost all parameters of interest across levels of the manipulated variables. Both R/JAGS and BLImP/Mplus encountered convergence issues and returned inaccurate parameter estimates when imputing and analyzing dichotomous data. Follow-up studies showed that JAGS and BLImP returned similar imputed datasets but the choice of analysis software for MLM impacted the recovery of accurate parameter estimates. Implications of these findings and recommendations for further research will be discussed. / Dissertation/Thesis / Doctoral Dissertation Educational Psychology 2016

In this study I addressed three problems related to how sandhi-variation, the adjustments made by speakers to the speech stream, filters comprehension for second language listener processing. The first was the need to better understand proficiency problems encountered by L2 listeners as they decode the speech stream with the phonological features of sandhi-variation, elision and assimilation, by investigating the item difficulty hierarchy of the phenomena. The second was the scarcity of research on aural processing abilities of second language learners in relation to their understanding sandhi-variation in aural texts. The third concerns the lack of research investigating links between learners’ backgrounds and their ability to handle listening texts, especially variations in the speech stream in target aural texts. The purpose of this study was threefold. My first purpose was to investigate the item difficulty hierarchy of sandhi-variation types that learners have in relation to L2 listening proficiency. My second purpose was to evaluate links between aural input containing elision and assimilation and second language aural processing, to provide insight into how learners deal with sandhi-variation as they process such input. My third purpose was to investigate through the use of interviews the aural input that participants have encountered prior to the interventions of this study, to help explain which types of aural input can facilitate intake. Twenty-five first- and second-year Japanese university students participated in the current study. The participants completed a series of instruments, which included (a) a Test of English as a Foreign Language Paper-Based Test (TOEFL PBT), (b) a Listening Vocabulary Levels Test (LVLT), (c) a Modern Language Aptitude Test–Elementary (MLAT-E), (d) a Pre-Listening in English questionnaire, (e) an Elicited Imitation Test (EIT), and (f) a Background and Length of Residency interview. The EIT was used as a sandhi-variation listening test with two component parts (i.e., elision and assimilation) and two sub-component parts (e.g., two different utterance rates), using elicited imitation. Finally, the participants were interviewed about their language backgrounds to gauge their understanding and feelings about English. An empirical item hierarchy for elision and assimilation was investigated, along with the determinants of the hierarchy. Overall, the tendency was for items with elision and assimilation to be more difficult. Results also indicated that the two input rate variables combined with elision and assimilation affected the non-native participants’ listening comprehension. Moreover, the strength of the relationship between two measures of the participants’ language ability, proficiency and aptitude, and their comprehension of items with and without the phonological features of elision and assimilation, were investigated. The results confirmed a positive relationship between language aptitude as measured by the MLAT-E and the comprehension of the phonological features of elision and assimilation. Finally, the results indicated that there were no significant, positive correlations between English language proficiency scores and both the Pre-Listening Questionnaire, which measured the participants’ feelings about second language listening, and the Background and Length of Residency Interview. More research needs to be conducted to determine how learners’ backgrounds are related to listening comprehension in order to better prescribe aural input in second language listening classrooms. / Teaching & Learning

The Upward Bound Math and Science (UBMS) program at Temple University (TU) seeks to guide first-generation, low-income high school students from Philadelphia in their quest to obtain postsecondary STEM degrees that lead to related careers. This study aims to evaluate the efficacy of the program by analyzing its students’ performance in light of its own goals and also in view of those of its sister program, The Upward Bound (UB) program. Primarily, this study uses Confirmatory Program Evaluation to ascertain which programmatic elements predict outcomes associated with indicators of STEM interest. Program leaders in TU’s College of Education and Human Development provided data for the study. Data came from 2015–2021 annual performance reports, containing information on 374 former UBMS student participants as well as 483 former UB students. Findings indicate that UBMS students pursued postsecondary enrollment at a significantly higher rate than UB students. Results also suggest that UBMS students’ grade level upon program entry was a significant direct negative predictor for how long they were affiliated with the program. In addition, students’ grade level upon program entry was a direct positive predictor for whether they obtained a rigorous course of study. Community service involvement was a significant negative predictor for whether students completed advanced math courses. None of the predictors yielded a significant effect in either of the two outcomes by way of a third, mediating variable. Implications of the study suggest that alliances between principal investigators, program administrators, and statisticians—ultimately through mixed-methods approaches—may offer valuable insight regarding the evaluation of UBMS and programs like it. / Educational Leadership

The purpose of this study was to analyze the impact of the socioeconomic factors of parent education level and family income on the academic achievement of students of Hispanic and white ethnicities. Scaled scores from the 2009 administration of the California Standards Tests in English language arts and mathematics and matched demographic information for 18,000 second through fifth grade students from six school districts in the San Joaquin Valley constituted the data source for this study. Multiple regressions were the primary statistical test used to analyze the data. The results showed a statistically significant gap in achievement between Hispanic and white students. After correcting for socioeconomic status and students of limited English proficiency, a residual achievement gap of roughly 0.1 of a standard deviation remained between white and Hispanic students. Further analysis showed no gap at low socioeconomic levels and a widening discrepancy in scores with increasing family income and parental education levels. These results may be indicative of differential expectations for white and Hispanic students. Additional testing for the effects of school-wide variables found a small negative impact on student achievement for schools with high average parental education levels, possibly caused by stronger interventions at schools with low average parental education.

Since the late 1800s, testing has been a part of education. Supporters claim testing encourages objectivity in assessments of learning. Critics argue testing creates bias, restricted teaching methodology, and restricted curriculum. The No Child Left Behind Act of 2001 requires school accountability that has increased the push for standardized testing. High school students present a unique challenge for administrators on the low-risk California Standards Tests (CSTs). The CSTs in mathematics are complex in reference to student composition, range of courses tested, and sequence of previous math courses taken by the student. This complex nature of the CSTs in mathematics makes test preparation decisions difficult for the site administrator. Schools have developed test preparation practices in the hopes of improving student performance on these standardized tests. Some of these practices require a great deal of time, energy, and money to create and execute. The administrator must use these precious resources judiciously. Therefore, there is a need for further investigation of the criteria used by site administrators for selection of test preparation practices for CSTs in mathematics. Using qualitative analysis, this study researched the criteria used by the high school site administrators to govern their decision of which test preparation practices for CSTs in mathematics to establish at their school. This study included 18 administrators at 15 schools in the southern San Joaquin Valley in California that met their Academic Performance Index growth targets for 2006.