The Lake Wobegon effect in student self-reported data

Abstract

For over a decade, Garrison Keillor assured listeners of American Public Radio's A Prairie Home Companion that in the women are strong, all the men are good-looking, and all the children are above average (Judith Yaross Lee, 1991). Although we can offer no explanation for the women's strength and men's appearance, we may be able to shed light on the reports of exceptional achievement of Wobegon's children. Our research indicates that students tend to overstate their academic accomplishments, and that belowaverage students are less likely to report their achievements at all. This produces a Lake Wobegon effect in student selfreported data. Overstated achievement may produce biased estimates of the relationship between achievement and educational inputs, if overstatement is correlated with achievement. In economic education, the use of student self-reported data for empirical analysis is not uncommon. Both the National Assessment of Economic Education data base and the third edition of the Test of Understanding in College Economic data base use some self-reported data. In addition, some economic-education researchers develop their own data sets that include student self-reported data. Systematic reporting biases in data collection also create problems for empirical analysis. Nonrandom reporting has been documented as a potential problem in economic education. Students are more likely to report grading errors on examinations when the correction of these errors results in raising their grades (Dale 0. Cloninger and Robert Hodgin, 1986). Sample selection bias resulting from using an unrepresentative sample has been shown to result from data lost between pretests and posttests (William E. Becker and William B. Walstad, 1990), students' failure to take standardized tests (Claude Montmarquette and Rachel Houle, 1986), and selection procedures used to place students in experimental and control groups (John J. Siegfried and George H. Sweeney, 1980).

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 110 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.