Because you are writing this column as Lennon and McCartney, that is without telling who actually did it, it is strange to use authorial “I” inside the column. Also, if you expected quality comments, you’ve got them!
Just that when we deride sensationalism so much around here, a hyperbolic lead in to your column sounded somewhat ironic. To start an article about fake data with the (perhaps trivial) fake data point that “we are awash in fake data.”
“It’s been nearly 20 years since the last time there was a high-profile report of a social science survey that turned out to be undocumented. I’m referring to the case of John Lott, …”
This is inaccurate, or at least disingenuous. It wasn’t based on survey data, but Michael A. Bellesile’s 2000 book _Arming America: The Origins of a National Gun Culture_, which claimed that private gun ownership was uncommon in antebellum America, was based on falsified data. How come you never mention this case?
I wasn’t talking about fraud in general, I was specifically referring to social science surveys. The examples of Hauser and Bellesiles and Stapel and others are interesting, and represent different sorts of fraud, but I wouldn’t call them surveys, at least not in the same way that Lott and LaCour reported surveys that were intended to be representative of a larger population. There’s lots of fraud out there to be discussed; I just happened to be talking specifically about this particular issue of reported survey results with no documentation.
Because you are writing this column as Lennon and McCartney, that is without telling who actually did it, it is strange to use authorial “I” inside the column. Also, if you expected quality comments, you’ve got them!
Fake data can be fun! And sometimes it can be used to invalidate models which use real data: http://zacharydavid.com/2014/05/know-thy-model-specificity-and-the-importance-of-using-fake-data/
Faking can be a coast-to-coast phenomenon. Darsee at Harvard
http://articles.latimes.com/1987-04-29/news/mn-1461_1_biomedical-research/3
and Slutsky at California
http://jama.jamanetwork.com/article.aspx?articleid=376306
A rather famous criticism of a Darsee publication is
“In one family tree, for example, a 17-year-old was listed as the father of children aged 8, 7, 5 and 4.”
And then there is the Infuse case of Timothy Kuklo who not only forged the data but also forged his coworkers
http://www.nytimes.com/2009/05/21/business/21surgeon.html?_r=0
Isn’t declaring that “we’re awash in fake data” a bit of an exaggeration? Are things really that bad?
Rahul:
Call it a copout, but . . . I didn’t write that subhead.
Andrew:
It didn’t matter for the substance of your piece.
Just that when we deride sensationalism so much around here, a hyperbolic lead in to your column sounded somewhat ironic. To start an article about fake data with the (perhaps trivial) fake data point that “we are awash in fake data.”
“Are things really that bad?”
ye old known unknowns & unknown unknowns
we have a sample of faked data from a population of published science studies
how good are we at sampling for faked data, in your opinion
“It’s been nearly 20 years since the last time there was a high-profile report of a social science survey that turned out to be undocumented. I’m referring to the case of John Lott, …”
This is inaccurate, or at least disingenuous. It wasn’t based on survey data, but Michael A. Bellesile’s 2000 book _Arming America: The Origins of a National Gun Culture_, which claimed that private gun ownership was uncommon in antebellum America, was based on falsified data. How come you never mention this case?
Anon:
I wasn’t talking about fraud in general, I was specifically referring to social science surveys. The examples of Hauser and Bellesiles and Stapel and others are interesting, and represent different sorts of fraud, but I wouldn’t call them surveys, at least not in the same way that Lott and LaCour reported surveys that were intended to be representative of a larger population. There’s lots of fraud out there to be discussed; I just happened to be talking specifically about this particular issue of reported survey results with no documentation.