Assessing Linguistic, Mathematical, and Visual Factors Related to Student Performance on the Texas Assessment of Knowledge and Skills, Eighth Grade Mathematics Test.
Description: The No Child Left Behind Act and National Council of Teachers of Mathematics' Principles and Standards both had a significant impact on the format and content of the Texas Assessment of Knowledge and Skills (TAKS) math test. Content analysis of the 2004 TAKS eighth grade math test identified the prevalence of linguistic complexity, mathematical rigor, and visual presentation factors and explored their relationship to student success on individual test items. Variables to be studied were identified through a review of literature in the area of reading comprehension of math word problems. Sixteen variables of linguistic complexity that have been significantly correlated with student math test performance were selected. Four variables of visual presentation were identified and ten variables of mathematical rigor. An additional five variables of mathematical rigor emerged from preliminary study of the 2003 TAKS math test. Of the 35 individual variables, only four reached a significant level of correlation with the percent of students correctly answering a given test item. The number of digits presented in the problem statement and number of known quantities both exhibited a significant positive correlation with the dependent variable. The number of times a student had to perform a multiplication operation had a significant negative correlation with the percent of correct responses, as did the total number of operations required. Stepwise regression of these four variables revealed total number of operations and known quantities to be the best combination of predictors of correct responses. When grouped in categories by problem type and compared, items involving mathematical reasoning but no mathematical operations had a significantly higher percentage of correct responses than those requiring at least one operation. Further categorization revealed problems involving applications only (without computation) associated with the highest levels of correct responses, followed by those involving only computation. Items requiring both applications ...
Date: August 2005
Creator: Norgaard, Holly Luttrell
Partner: UNT Libraries