From the faculty lounge: False Positives

Punch line: Sometimes, published academic research results are flat out wrong.  Hmmm.

Excerpted from HBR’s Daily Stat: Researchers Can Easily “Prove” False Findings
 
Using legitimate statistical analyses, researchers were able to show in an experiment that participants were nearly 1.5 years younger after listening to the Beatles’ “When I’m Sixty-Four” than after listening to a song that comes with the Windows 7 operating system …

… an obviously ridiculous finding that demonstrates how easy it is for research to yield “false positives,” say Joseph P. Simmons and Uri Simonsohn of The Wharton School and Leif D. Nelson of UC Berkeley.

Too often, researchers aren’t aware of the high likelihood of finding false evidence, and the pressure to publish leads scientists to convince themselves of the validity of their findings, the authors say.
 
Source: False-Positive Psychology : Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant

Surprise, surprise, surprise.

>> Latest Posts

Overheard in the faculty lounge … re: the Target flap.

By now, everybody has heard how Target mines data on shoppers to ID when they’re approaching life events — e.g. having a baby — that make them “vulnerable to marketing initiatives.”

The reaction of many marketers seems to be: “why aren’t we doing that?”

The reaction of shoppers is predictably negative: “Invasion of privacy”, “manipulative”, “creepy”.

The reaction in the faculty lounge is interesting.

Background: a branch of marketing studies consumer behavior … how and why consumers think and act … why they pick one brand over another, etc.

There seems to be concern among some academic CB researchers that their findings are  being hijacked by evil profiteers, to the disadvantage of the masses:

Consumer behavior research clearly helps the stores in the “attack” on the consumer. Does CB help in the development of the “defense” of the consumer?

One colleague sought to allay any pangs of guilt:

The “consumerism” defense is that the findings can be used to benefit both producers and consumers.

Any way, as [a famous consumer researcher] used to argue “the effects we study are so small in the real setting that any harm done is minimal.”

Now, that’s a rallying cry for you …

>> Latest Posts