In an analysis of education articles published in The New York Times, Washington Post, and Education Week, Holly Yettick of the University of Colorado at Boulder finds that any given think tank report was substantially more likely to be cited than any given study studies produced by a university.
Her study of 864 articles shows that
1. Education Week most often cited university-based research, while The New York Times and The Washington Post most often cited research produced by governmental entities.
2. Although university and government sources were cited more often, a higher percentage of reports produced by advocacy-oriented think tanks were cited by both types of publications. Universities produce 14 to 16 times more research than think tanks, but the three publications only mentioned their studies twice as often as think tank reports.
Given these findings, Yettick recommends that education reporters and editors adopt the following guidelines when writing about educational research:
• Expand your source list. The findings of this study suggest that think tank research is over-represented in media coverage. Unlike think tank employees, university professors generally lack the incentives and resources to conduct public relations campaigns involving outreach to journalists. However, many would like their research to reach the public. Like their science- or medical reporting peers, education reporters should consult peer-reviewed research and cultivate university researchers, who should be able to recommend major, peer-reviewed studies in their fields.
• If you do decide that a think tank study merits recognition, do your own quality control. Vet reports before publishing. Most research reports will not lose news value during the time taken to verify their soundness. A good method of conducting such verification is to consult with a trustworthy person with expertise in research design and statistics. … In addition, consult subject matter specialists, ideally those who have read the report. If the reporter is only able to consult subject-matter experts who have not read the report, note this in the article, helping readers understand that the study’s findings should be taken with caution until experts have had time to fully review the results.
• Include full disclosure. Regardless of who produced the study, the article should link to the full report so readers can judge for themselves. Non-peer reviewed research should also be labeled as such.
Update: Read Kevin Carey’s response in The Chronicle
Posted by paul baker 



