Two academics think now-disgraced UCLA grad student Michael LaCour faked the data in yet another study, this one on whether people get their news primarily from the “echo chambers” of television news.
In recent days it has come to light that LaCour likely faked data related to an explosive survey he claims to have conducted of same-sex marriage opponents in California. The paper, published in the presitigious journal Sciecne, has since been removed.
George Mason University Professor Tim Groseclose became suspicious when he was asked to examine another LaCour paper, “The Echo Chambers are Empty: Direct Evidence of Balanced, Not Biased Exposure to Mass Media.” This “unpublished working paper” examines the news “diet” of Democrats and Republicans and concludes, according to Groseclose, “voters do not, primarily, get their news from ‘echo chambers’.”
What LaCour claimed to have done is to compare buzz phrases — “death tax” vs. “estate tax” — used by Congressional Democrats and Republicans and then compared the frequency of use across various media outlets.
On a post at the website Ricochet Groseclose goes into detail about why he was almost certain LaCour had faked this study. Much of it is so complicated that the average graduate student in the field of political science would not understand it. And this is where LaCour may have gotten into trouble. Groseclose says LaCour didn’t understand the statistical models he was cribbing from. In fact, according to Groseclose, LaCour did a cut and paste job of someone else’s work in explaining his statistical modeling.
What’s more, it appears that LaCour cribbed some of his data from Groseclose’s own data file, which Groseclose had provided to him a few years ago. This would not have been a problem except that LaCour said he was using a different data set.
As Groseclose explained to Breitbart News, “The problem is that his paper claims to estimate results from other data (loaded-phrase data), whereas the code he writes seems to use my data. This means that he couldn’t have used that code to derive the results that he reports. Either he sent Martin the wrong code or, what I think actually happened, he simply faked the results he reports then when Martin asked him for code he sent the closest substitute he could find (which I believe came from a homework problem one of his professors wrote).”
Emory University’s Gregory J. Martin published a brief paper on May 29 arriving at the same conclusions as Groseclose. Martin wanted to replicate LaCour’s conclusions and requested LaCour’s data set and after trying and failing to run it concluded, “Simply put, despite using the same dataset of news show transcripts and implementing the same method described in his paper my results were not at all comparable to LaCour’s.”
Martin says LaCour’s results “are very likely fabricated.”
Besides the highly technical and nearly indecipherable methods of statistical measurement and analysis that LaCour got wrong, he also made some glaring errors. For instance, he claims to have used the “UCLA Closed Caption (LACC) archive of news show transcripts.” However, for the year he claimed to have measured, 28 of 60 shows were not even the UCLA database.
Groseclose considers this the biggest scandal ever to hit the field of political science, and he thinks more scandal are coming and for the same reason LaCour got caught. The methods are simply too complicated even for many graduate students to understand.
He says, “Over the past five or ten years I have have noticed more and more papers written by young political scientists (grad students and assistant professors) that claim to use extremely fancy and complex statistical techniques, yet the authors do not seem to fully understand the techniques they claim to use. Their descriptions of their statistical methods are often as opaque as the LaCour index I discuss above.”
He says within political science there are probably about half a dozen additional Michael LaCours — researchers who are perceived to be solid and talented scholars yet have build that perception partly upon faked results.”
Groseclose predicts that LaCour will not be awarded a Ph.D. from UCLA and his job offer from Princeton, set to begin this summer, will be withdrawn.
Follow Austin Ruse on Twitter @austinruse