Little published work has appeared on the teaching of statistics in economics. This lack of attention is surprising since John Seigfried and James Wilkinson (1982) found that nearly 85 percent of the 546 schools surveyed required at least one course in statistics for an undergraduate major in economics, and approximately 50 percent of the departments of economics offered their own course. After introductory economics, only intermediate macroeconomics and money and banking have higher enrollments than statistics. This paper focuses on the goals and objectives of the introductory economics statistics course. It also gives attention to the more advanced econometrics courses undergraduates may elect to take.' It emphasizes the quantitative skills all undergraduate majors in economics should possess irrespective of their career plans. As pointed out by W. Lee Hansen (1986), it is much easier to talk about what should be in a course than to define the student competencies to be gained. The latter requires the definition of student activities while the former treats the student as a passive recipient. If students are to do more than regurgitate definitions, duplicate proofs, or perform repetitive computations, then what is expected from them must be specified. Whether an instructor covers something or whether students can actually do something with what is covered are two different issues. This paper focuses on what students should be expected to do, not on what should be in economics statistics courses.

Following the completion of microand macroeconomics principles courses, for which the majority of students are enrolled to fulfill requirements for other majors, economics majors take two intermediate micro and macro courses, a course in statistics/econometrics, and some field courses that may or may not include more quantitative methods (John Siegfried et al., 1991). In a national survey, Michael Watts and I found some differences between the way statistics and econometrics courses are taught and the way the other undergraduate economics courses are taught (Becker and Watts, 1996). In particular, problem sets are used more in statistics and econometrics than in other undergraduate economics courses. Curiously, however, those applications are not based on events reported in newspapers, magazines, and journals that economists read. How timely and relevant can problem sets be if they are not documented in current events? Ideally instructors set problems raised by their own research and consulting; problems students can expect to see on their jobs. After all, the rationale for teaching statistics and econometrics outside a mathematics department rests on a belief that there is something special about economic analyses. That is, economists' use of statistics is tied to the issues they face. Although the calculation of a mean and a median, for example, is the same in medicine and economics, a discussion of the average duration of economic expansions since World War II is more pertinent to those majoring in economics than is a discussion of average blood pressure or average time to dementia with mad-cow disease. The importance of economic tleory is often lost when mathematicians attempt to make situations real, as seen for example in the Chance Course (J. Laurie Snell and John Finn, 1992), where a potpourri of statistical applications are presented with no disciplinegrounded analyses. To teach students to apply the tools of statistics to actual situations and data encountered by economists, there is little justification for examples involving the drawing of balls from urns, flickng of spinners, tossing of coins, or contrived card and dice tricks. Yet these methods of generating data continue to be found in the activity-based teaching and assessment Discussants: William Greene, New York University; Robin Lumsdaine, Brown University; Kim Sosin, University of Nebraska-Omaha.