Volume 1 - Issue 1 - DBU Journal for K-12 Educational Leadership - Page 29

Journal for K-12 Educational Leadership 27 between ELLs and English fluent students. More than 40 years after the first Supreme Court case, the achievement gaps still exist. In 2013, ELLs scored 41 per- centage points lower in reading and 47 percentage points lower in writing on Texas’ state achievement test than their English fluent peers at the high school level (Texas Educa - tion Agency, 2014). There are limited-English and non-En- glish speaking students in secondary classrooms who are not learning at desired rates because they do not sufficiently understand academic English. Often, teachers also strug- gle to understand the language of the ELLs. When stu- dents and teachers do not understand one another, effective teaching cannot occur and “experts on…language minority children… agree that effective teaching is critical to student learning” (Calderón, Slavin, & Sánchez, 2011, p. 118). Teach - er knowledge about students and teacher attitudes toward students have long been associated with student achieve - ment. Teacher knowledge about second language acquisi - tion (SLA) and a teacher’s attitude toward ELLs can affect the achievement of these students. Research Questions Theover–archingquestionof thecurrent studywas:Howwell doteachers’knowledgeandattitudesaboutELLsandSLApre - dict ELLs’ achievement on the STAAR test? Specifically, the currentstudyexaminedthefollowingfourresearchquestions: Research Question 1 (RQ1) Do teachers’ years of expe- rience and hours of ELL professional development predict ELLs’ achievement on the STAAR and STAAR EOC tests? Research Question 2 (RQ2) Do secondary teachers’ knowledge about ELLs and SLA, as scored on the Con- tent-Area Teacher Survey , predict ELLs’ achievement on the STAAR and STAAR EOC tests? Research Question 3 (RQ3) Do secondary teachers’ at- titudes about ELLs and SLA, as scored on the Content-Area Teacher Survey , predict ELLs’ achievement on the STAAR and STAAR EOC tests? Research Question 4 (RQ4) When taken together, do secondary teachers’ knowledge and attitudes about ELLs and SLA, as scored on the Content–Area Teacher Survey , pre- dict ELLs’ achievement on the STAAR and STAAR EOC tests? Results & Findings Hierarchical multiple regression was used to compare teach - er survey results with ELLs’ state achievement test scores to determine if teacher knowledge and attitudes about ELLs and SLA were statistically significant in predicting student achievement on Texas’ state achievement test. Fifty-seven secondary teachers and 58 ELL students participated in the study. Table 1 reveals a mean of 8.8 years and a mean of 14.7 hours for the two covariates, teacher’s years of experience and teacher’s hours of ELL training, respectively. (See Table 1 on page 30.) Table 2 also reveals the mean of the two predictor variables, teacher knowledge and teacher attitudes toward ELLs and SLA. The mean for teacher knowledge was 211.8 and the mean for teacher attitude was 400.9. A high score of 300 with a low score of 60 was possible for each student in teacher knowledge and a high score of 450 with a low score of 90 was possible for each student in teacher attitude. (See Table 2 on page 31.) A teacher’s years of experience and hours of ELL and SLA professional development, the two covariates or nui - sance variables, were entered into the regression equation first to establish a baseline beta-weight. Once the baseline beta-weight was established, teacher’s knowledge about ELLs and SLA was entered into the equation to determine its predictive ability, or change in R 2 , of ELLs’ STAAR scores. Finally, teacher’s attitude about ELLs and SLA was entered into the regression equation to determine its predictive abil- ity, or change in R 2 , of ELLs’ STAAR scores. When taken individually, teachers’ years of experience was a strong predictor of ELLs’ STAAR scores, but teachers’ hours of ELL professional development was not a strong predictor of ELLs’ STAAR scores. However, as model 1 in Table 2 shows, the two covariates, when taken together, have a large effect in practice (Cohen, 1988), accounting for 34% ( R 2 = .340) of the variance in STAAR scores. An ANOVA was conducted to test whether the regression model, with both covariates included, significantly predicted ELLs’ STAAR