I think that word does not mean what you think it means

[aroraborealis]

Today, in Wow, But People Suck:

A Silenced Drug Study Creates An Uproar

Within the company, meanwhile, officials explicitly discussed misleading physicians. The chief of a team charged with getting articles published, John Tumas, defended "cherry-picking" data.

"That does not mean we should continue to advocate" selective use of data, he wrote on Dec. 6, 1999, referring to a trial, called COSTAR, that also produced unfavorable results. But he added, "Thus far, we have buried Trials 15, 31, 56 and are now considering COSTAR."

Can you still call it data after it's been cherry-picked?

Comments

[zzbottom.livejournal.com]

It's like going to the doctor and getting your health graded on a curve. "We used to think that a 275 cholesterol count was dangerously high, but since the national average is now 235, we're just going to tell you it's slightly elevated."

[ectropy.livejournal.com]

Sure, it's still data, but it's not bias-free information.

[water-childe.livejournal.com]

there seems to be a rash of these sorts of stories being reported in the last few months. cherry picking is bad enough, but in some recent cases the data is sometimes mostly fiction. i'm still pissed off about the deal with andrew wakefield and the false data he invented in regards to children with children with autism and mmr vaccination.

http://www.huffingtonpost.com/2009/02/08/doctor-andrew-wakefield-w_n_165033.html

then there is dr. ruben, at bayside medical center. it's come to light that he falsified data, and even 'invented' patients.
his studies have had a great deal of influence as to how doctors treat surgery patients for pain. 21 published papers. all false. 13 years. it's sick.

http://www.masslive.com/news/index.ssf/2009/03/baystate_medical_center_probe.html

[water-childe.livejournal.com]

forgive me my bad typo.
children with autism and mmr vaccination.
children with children and mmr vaccination makes it sound like the study deals with pregnant child brides and their autistic offspring.

[dr-memory.livejournal.com]

Can you still call it data after it's been cherry-picked?

Sure, as long as it's more than a single datum. :)

[concrete.livejournal.com]

Have you read the wonderful and very short book called "How To Lie with Statistics" ?

[randysmith]

AyyyyYuck!

You know, I can't help but wonder if the executives who do this really understand what they're doing. If you understand statistics, you know that burying negative data invalidates statistical results from positive data, and hence is morally equivalent to lying about your drug. But I get the impression that not many people understand statistics.

The executives in my company (I believe) don't lie, but put the best spin on the truth that they can manage, and most of the time I think that's an honorable thing for them to do. Maybe the people at these companies think of burying bad data as another way to put a good spin on the honest data they have?

It does bring up the amusing idea of sentencing these people to jail time, which is waive-able if they take and get a good grade in a reasonably rigorous statistics course :-} :-J.

[snowy-owlet.livejournal.com]

As of last September, the Clinical Trials Database (ClinicalTrials.gov) has required reporting of results, specifically to counteract this problem.

STABBITY STAB!

[drwex]

Not that I'm biased or anything, but man this sort of crap makes me stabbity.