NotesFAQContact Us
Search Tips
ERIC Number: ED170306
Record Type: Non-Journal
Publication Date: 1976-Dec-8
Pages: 60
Abstractor: N/A
Reference Count: N/A
Prediction of Test Scores: California.
Wise, David A.
Census data were used to predict average district scores on standardized achievement tests for grades one, two, three, six, and twelve, in 716 California school districts. The following census variables were used: education, income, college, poverty (families); poverty (children); professional occupations; housing; Spanish surnames; and Black. Results are reported both as absolute scores (average number answered correctly) and as transformed scores (a standardized score which allows grade level comparisons). Predictions were presented in a series of tables. The first predictions included only a small number of census variables, beginning with a measure of poverty--with more variables added in subsequent formulations. The reason for following this format is to give some idea of the gain in "predictive power" to be had by including additional variables and to allow alternative specifications for simulation. Results indicated that the predictive power of reading test equations, based on average education and median income of district residents, was not improved substantially by adding other educational and income variables--these additional variables did improve the predictive power of the mathematics test equations. Ethnic variables increased the predictive power considerably. In general, there was no advantage to using transformed scores. (The tables, comprising over one half of the document, are appended). (CP)
Publication Type: Reports - Research; Numerical/Quantitative Data
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Identifiers - Location: California