In DepthSCIENTIFIC INTEGRITY

Survey fraud test sparks battle

Science  04 Mar 2016:
Vol. 351, Issue 6277, pp. 1014
DOI: 10.1126/science.351.6277.1014

You are currently viewing the summary.

View Full Text

Summary

At a meeting on survey data fabrication in Washington, D.C., last week, Michael Robbins and Noble Kuriakose presented an update on their statistical test that has been roiling the survey research community for the past year. When they apply the test to more than 1000 public data sets from international surveys, about one in five of the surveys fail, indicating a high likelihood of fabricated data. At the meeting, they debuted an analysis, focusing on 309 international studies funded by the Pew Research Center, that found a failure rate of 30%. "Robbins and Kuriakose have uncovered a massive problem," says Michael Spagat, an economist at Royal Holloway, University of London, who has investigated high-profile cases of possible survey data fabrication in war zones. But Pew officials dismiss the test, saying it is prone to false positives. The organization has gone so far as to request that Robbins and Kuriakose desist from publishing their original analysis, which is now in press.