Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
INVESTIGATION OF AN ERROR ANALYSIS METHODOLOGY FOR THE STUDY OF SEX DIFFERENCES IN MATHEMATICS ACHIEVEMENT
Abstract
This research investigated a methodology for examining sex differences in mathematics performance through multi-level analysis of error responses on the ITBS Mathematics Problem Solving Test for grades three, six, and eight. Item distractors were ranked according to their desirability as a response choice based on their error severity by three mathematics education experts. Only those items receiving a judges' consensus on distractor ranks were included in the analyses of error responses. It was expected that sex differences were unlikely for grade three, more likely for grade six, and most likely for grade eight. T-tests on mean total-number-correct scores over all test items for each grade indicated significant differences favoring girls for grades six and eight. These findings led to an expectation of sex differences in subsequent analyses of error responses to items with distractor rankings. This expectation was not substantiated. Separate ANOVAs by grade were performed on error response proportions for items with desirability ranks in combination. For grades three and eight, factors in the analyses were sex, distractor desirability, and item difficulty; for grade six, only sex and distractor desirability were included. That no significance attributed to sex was observed for any grade indicated no differential tendency overall for one sex to select more desirable distractors than the other. Individual items were subjected to contingency table analyses to uncover differential error response patterns by sex. Significant (p < .01) differences were found for only nine of the tested items (16-17% for each grade). One group goodness of fit tests of these items indicated distributions were non-random for both sexes. Items were examined in respect to skills objective, context, gender of actors in the word problem, difficulty, and types of errors across grades. No patterns emerged to explain significant findings for these particular items. Although the findings suggest that this methodology may not be highly useful, they do suggest that research employing tests specifically designed with error analysis in mind or using more traditional error analysis methods may have potential benefit for studying sex differential performance in mathematics.
Subject Area
Educational psychology
Recommended Citation
SMITH, ELIZABETH PRINCIPE, "INVESTIGATION OF AN ERROR ANALYSIS METHODOLOGY FOR THE STUDY OF SEX DIFFERENCES IN MATHEMATICS ACHIEVEMENT" (1984). ETD collection for University of Nebraska-Lincoln. AAI8423829.
https://digitalcommons.unl.edu/dissertations/AAI8423829