UBC Theses and Dissertations
Cognitive and noncognitive correlates of statistics achievement Selkirk, Sheena Ann
This thesis is a report of an attempt to account for the observed variance in achievement in introductory statistics, within a population of graduate students in Education. Measures of twelve independent variables, selected on the basis of the literature in the area, were collected-from volunteer subjects during a class session of each of four classes at the Universities of Alberta, British Columbia and Victoria. The variables included measures of verbal comprehension, logical reasoning, spatial visualization, number facility, and concept learning. Measures of attitude toward statistics, of attitude toward research and of academic self-concept were collected, as were measures of achievement motivation, of specific motivation and of perceived utility. The twelfth independent variable was experience, as measured by the number of mathematics or statistics courses previously completed. The measure of the dependent variable, statistics achievement, was the score obtained on a one-hour common examination. The original intent was to investigate the determinants of statistics achievement through the application of Bentler's "causal modeling with latent variables" (LVM) approach. Unforseen delays in administering the measures of the independent variables at the University of Alberta resulted in statistically significant differences between the British Columbia and Alberta pools of data. Consequently, the research questions were redeveloped so as to exclude consideration of the LVM approach and any exploration of potential latent variables from the analysis plan. The method of analysis was limited, then, to the examination of zero-order correlations and to the computation of exploratory forward regressions. The British Columbia and Alberta samples were analyzed separately. Because of the exploratory nature of the analysis, the 0.10 level was adopted as the criterion of statistical significance. Correlational analysis of the British Columbia sample revealed that experience, verbal reasoning, spatial visualization and concept learning significantly correlated with achievement. Two clusters of intercorrelations (cognitive-experiential and attitudinal-motivational) were identified. Only the first cluster was related to statistics achievement. Application of stepwise regression revealed that only past experience accounted for a statistically significant proportion of the variance in statistics achievement. The set of 12 variables taken together accounted for 55.5% of the variance. Similar analysis of the Alberta data revealed that spatial visualization, number facility, concept learning, attitude toward statistics and attitude toward research significantly correlated with statistics achievement. Cognitive-attitudinal and attitudinal-motivational clusters of independent variables were identified in the correlation matrix. Both clusters appeared to be related to statistics achievement. The results of a stepwise regression revealed that 54.9% of the variance in achievement was accounted for by the set of 12 independent variables. Three variables (concept learning, attitude toward statistics and spatial visualization) contributed significantly to the explained variance in statistics achievement. When the two samples were subjected to exploratory regression analyses, the order of entry of the independent variables was not similar in the two samples. The results were discussed in terms of the findings reported in the literature and of the possible effects of the different administration times of the tests. The thesis is terminated by a presentation of some suggestions for future research.
Item Citations and Data