Sheila Dow has gathered several economists to review Card and Krueger's important work Myth and Measurement: The New Economics of the Minimum Wage, and what is more important than what any of these individual reviewers has to say is the collective impression left by the reviewers: the politics of research play a far larger role than do the numbers themselves. Finis Welch is left calling for an entirely new set of standards to govern the process of economic research and interpretation (Dow 2002).
Several of the reviewers in this article refer to the design of what Card and Krueger call their "natural experiments" designed to see how an increased minimum wage would affect employment. One of the primary arguments against an increase in minimum wage has been that employers would be forced to reduce their workforce numbers to accommodate the higher wage - or to raise prices to such a degree as to render the increased wage negligible. Card and Krueger find as their conclusion that this argument is not true - they conclude that the effect on employment of a higher minimum wage is almost nonexistent. Charles Brown (who, coincidentally, believes that a 10% increase in the minimum wage would decrease employment by 1%) criticizes Card and Krueger's methods, saying that their time window for measuring effects of an increased minimum wage is too short, and does not consider other factors, such as sunk and capital costs (Dow 2002).
By contrast, Richard Freeman, who views the minimum wage as one palliative for this "era with rising inequality," agrees with the use of a short time window, because "[e]mpirical analysis of quantitative responses to price incentives is most convincing over periods during which the structure of the economy, technology, and so on can safely be viewed as fixed"(Dow 2002). And so even one's opinion of statistical methodology, it seems, is related to one's prevailing opinion about the topic at hand.
If one wants to write a paper that has a certain result, the temptation will soon arise to allow the statistical analysis to provide only a particular answer. Paul Osterman makes note of an analysis that Card and Krueger conducted using past research. According to their study, not only have past economists decided not to publish papers about the minimum wage that would contradict the prevailing argument about the danger to employment numbers, but even within papers that were published, they have "simply [ignored] findings that were inconsistent with the point the authors wanted to make" (Dow 2002).
Of course, this manipulation and misuse of data is exactly what Finis Welch accuses Card and Krueger of doing in their own book: after dissecting Card and Krueger's studies and finding them too statistically insignificant to support a hypothesis, he asks, "Are these the kind of data that you would recommend as the basis for national policy"(Dow 2002).
In addition to flaws in the design and implementation of particular research methods, this selection of reviews also provides insight to the inherent biases that the reviewers themselves bring to