Analysis of a widely-publicised Science paper has found evidence that the data used the in the research was faked (PDF).
The study, by researchers Michael J. LaCour and Donald P. Green, claimed that a short conversation with a gay canvasser could persuade people in favour of gay marriage, significantly more than a conversation with a straight canvasser. The paper reported that the effect was found to last in 3-week, 6-week, and 9-month follow-ups, and could be passed on to other people living in the same household.
While working to replicate and extend the original research, David Broockman and Joshua Kalla uncovered problems with the data used by LaCour and Green. They were initially surprised at how high the response and re-interview rates were in the original paper’s data, and launched a small test run of their planned extension study.
Read 5 remaining paragraphs | Comments
