Predictive Validity of Formative Tests, statistics homework help
Assignment 2: Predictive Validity of Formative Tests
Suppose the state of Georgia requires all students enrolled in a 9th grade algebra class to complete a standardized end of course test (EOCT). The Georgia Department of Education further mandates that students’ scores on this test will be used in the annual evaluation of teachers who are responsible for the 9th grade algebra class. To assist teachers to better prepare their students for the EOCT through early identification of struggling students, a group of algebra teachers collaborate on the development of seven assessments suitable for use as quizzes in the target course. These seven formative assessment tools include items similar in wording, structure and format to those used on the EOCT.
This group of teachers shares their work with teachers across the state, and their quizzes are adopted by teachers in many districts. While some teachers use these assessments as part of their unit tests, others offer them as quizzes, extra credit homework assignments, and study guides for their own unit exams. After two years of use, the teachers using the seven quizzes would like to know if students’ performance on those assessment tools really is related to performance on the EOCT. Suppose you have been asked to analyze student data with the goal of evaluating the predictive validity of the seven formative assessments (referred to hereafter as the common assessments, or CA1, CA2, CA3, CA4, CA5, CA6, and CA7).
The population of interest for the study was all high school students in participating districts who successfully completed the 9th grade algebra class between fall 2015 and spring 2017. A census sampling approach was used. Because the data were collected by the school during regular classroom activities, and because no identifying information was gathered about participants, students and participants were informed the data would be used in this way but it was not necessary to obtain consent/assent. In May of 2017, the relevant data were downloaded from school data systems and submitted to you for analysis. You have cleaned the various data files and merged them into one for analysis. Detailed information about the items and coding of responses are presented in the Predictive Validity of Formative Tests Codebook document.
Research Questions
1. What are the characteristics of the sample and each subset of students included within the sample? (YUCEL)
4. What is the nature of the relationships between the EOCT subdomain scores?YUCEL)
7. To what extent, if any, do EOCT algebra and linear functions subdomain scores predict EOCT geometry subdomain scores?(YUCEL)
10. Explore the predictive validity related to the geometry subdomain of the
standardized EOCT :(YUCEL)
10a. To what extent, if any, is common assessment 4 a significant predictor of student performance on the geometry subdomain of the standardized EOCT?
10b. To what extent, if any, is common assessment 5 a significant predictor of student performance on the geometry subdomain of the standardized EOCT?
13. Write your own research question based on the data that are available to you. Analyze the data and present the findings as part of your report.(YUCEL)
- Variable 1 = SchoolTYPE. 1 = Public school, 2 = Charter school, 3 = Charter district, 4 = Private school
- Variable 2 = GENDER. 0 = female, 1 = male
- Variable 3 = race. 1 = Black or African-American, 2 = White, 3 = Other
- Variable 4 = domain1. EOCT domain 1 score: algebra and linear functions
- Variable 5 = domain2. EOCT domain 2 score: geometry
- Variable 6 = domain3. EOCT domain 3 score: statistics and probability
- Variable 7 = ca1. Formative course assessment 1 score
- Variable 8 = ca2. Formative course assessment 2 score
- Variable 9 = ca3. Formative course assessment 3 score
- Variable 10 = ca4. Formative course assessment 4 score
- Variable 11 = ca5. Formative course assessment 5 score
- Variable 12 = ca6. Formative course assessment 6 score
- Variable 13 = ca7. Formative course assessment 7 score
