The Impact of Teacher Leaders on Students' Achievement In Science
Teacher Leader Data
We administered pre- and post-assessments to TLs at
each of each two-week NCOSP Summer Academy (SA) to assess the development of their content
knowledge in physical science (SA 2004), life science (SA 2005), and Earth science (SA
2006), using the PET assessment (Goldberg, personal communication, 2004), the ATLAST Life
Science assessment from Horizon Research Inc., and the NCOSP Earth Science
Assessment. A year later, we administered delayed post-tests to examine TLs'
retention of the content knowledge. Also, on surveys at the end of each SA, TLs
responded to "retrospective" questions in which they rated their understandings before and
after the SA in a number of areas that reflect important aspects of effective instruction
(Bransford et al., 2000).
K-12 Student Data
Fifth, eighth, and tenth
grader students in Washington State take the science WASL each year. The tests
include 5-6 scenarios with approximately 35 items. The multiple-choice, short
answer, and extended response questions assess students' content knowledge in physical,
earth, and life sciences. Results are reported in terms of scale scores points and
the percentage of students meeting or exceeding the state standards for proficiency (i.e.
"proficient or above").
The NCOSP Internal Evaluation Team petitioned the Office of the Superintendent of Public Instruction (OSPI) to release WASL data from all NCOSP partnering districts. We received five years of WASL data beginning with the 2003-04 school year and ending with 2007-08 including results in Science, Math, Reading, and Writing for grades 3 through 10. The data collected represents 159 schools in 27 districts (84 elementary, 42 middle, and 29 high schools). The sample for each year includes approximately 5,000 students per grade.
The NCOSP Internal Evaluation Team compared science WASL scores for students within 27 partnering schools districts who were labeled as taught by a "NCOSP TL" or "non-NCOSP teacher". Additionally, we collected student demographic data regarding gender, ethnicity, and enrollment in special programs (e.g. Free/Reduced Lunch, special education, and transitional bilingual) to control for variables in our analyses that have been shown to affect student achievement. The four years of student achievement data from 2004-05 to 2007-08 were considered "treatment" years, while the data from the 2003-04 school year is baseline data. We examined the differences in student achievement in science for TLs vs. non-TLs for each school year, as well as changes in science WASL scores for NCOSP TLs over time using ANOVA models, regression models, and multi-level modeling (MLM).