Literature Index

A recent symposium on "Improving the Work Force of the Future: Opportunities in Undergraduate Statistics Education" was held to focus attention on the importance of undergraduate statistics education. The symposium and the approval of curriculum guidelines for undergraduate degrees by the Board of Directors of the American Statistical Association have done much to define the current state of undergraduate education in statistics and suggest directions for improvement. This article summarizes the activities leading up to the symposium and provides a brief summary of six papers from the symposium that have been published. The article concludes with a discussion of some of the outstanding issues that remain to be addressed.

Undergraduates at California State University, Chico's College of Agriculture do experimental research including data analysis. Desire to see if treatments differ, motivates students to learn inferential statistics. Students analyze data with ANOVA programs written for Microsoft Excel. These programs return an analysis with one or at most a few mouse clicks; they handle missing data - a problem in real world experiments. In statewide science competitions, our students routinely place first or second.

Children's understanding of what variables and relations are important in problem structures, and their use of these variables and relations in problem solving, were examined. One hypothesis suggests that knowledge of relevant solution variables is a prerequisite for encoding those variables, which in turn is a prerequisite for learning new strategies that use those variables. An alternative hypothesis holds that knowledge of relevant variables is an outcome, rather than a precursor, of efforts to invent new strategies. In the current studies, children between the ages of 5 and 13 years were given Piaget and Inhelder's (1975, The origin of the idea of chance in children, New York: Norton) two -set alternative choice probability problems. In Experiment 1, problem understanding was assessed by asking children to construct two-set problems that could test whether a learner understood how to solve a model problem type. In Experiment 2, understanding was assessed by asking children to modify model problems to make them harder for a learner to solve. In both experiments, children modified or reproduced only those properties of model problems used either correctly or incorrectly in solving the models. These results partially support both hypotheses, and suggest a mechanism by which problem solving knowledge develops.

This article uses a simple counter-intuitive example to point out a common misinterpretation of correlation.