Actually that should read..

I'm going to post one of three articles. Which would you prefer?

The first one.

Which would you prefer? The the first one or the third one.

The first one damnit.

Which would you prefer? The the first one or the second one.

THE FIRST ONE DAMNIT.

And tomorrow I'm giving you a choice between the second one and another (fourth) one.

(Thinks... This psychologist is really trying to force me to choose the second one... Well I'm not giving in. Not even if they put me in a box with a cat.)

The fourth one. Or else.

(Call me Ishmael.... there doesn't seem to be a way to set your name when commenting Anonymously)

We're not assuming that Egan and Chen know each other, Anonymous. Chen put Egan in the acknowledgments of the paper in question. I'd say that fairly removes doubt.

Yours forever,
Anonymous

Don't you mean: "Sorry New York Times, Cognitive dissonance has not been rejected"?

It seems that these studies conclude that they fail to reject cognitive dissonance...the Chen paper finds that some of these papers didn't actually fail to reject cognitive dissonance.

So we're left with a bunch of papers that didn't prove anything and one paper that fails to reject a hypothesis. [yawn]

Allan Pollock

It is difficult to carry on conversations with a bunch of people all named "anonymous." Or maybe it is only one.

I am not sure how people could have a preference for one object over another without knowing what they are. I am going to post one of two new articles today. Which would you prefer?

I think it is wrong to assume that Louisa and Keith knew each other because they are at yale. Keith works at SOM and Louisa for psychology. They are not in the same location
http://maps.google.com/maps?q=Department+of+Psychology,+yale,+new+haven&...

The blind experiment is interesting but still broken unless they follow Chet's advice. Lets assume the Null Hypothesis is: __measured__ preferences don't change by the act of choice but instead randomly fluctuate. If you repeat the experiment using the same methodology Chet has criticized you might mistakenly reject the null hypothesis because of random fluctuations in measured preferences.

Anonymous: People don't choose the unchosen option. They stick with what they previously picked, same as they do when they can see the objects in question.

I don't think I claimed Chen's email was a hatchet job. I brought up the fact that the recent study went unmentioned as an oddity. I am told that many other published studies are similarly immune to the criticism, but I don't know these other studies as well.

The point of the post, I had hoped, was to show that Chen brought up a potential problem with some studies, but it remains only a potential problem. So far there is no empirical evidence to support the claim.

So, at best, this is a potential problem with some, but not all, cognitive dissonance studies. But it is also possible that no cognitive dissonance studies actually have this problem.

Please try my web-based experiments

The unpublished study, from your description, doesn't seem that convincing. If they have no idea what they've chosen, then how can they rationalize their choice? My guess: the subjects don't give a damn, they pick the unchosen option for the novelty of it. This doesn't even begin to require the concept of dissonance.

I really don't think it's appropriate to suggest that this is a hatchet job by Chen. Your one piece of counter-evidence sounds flawed, and it's not published. Has this been shown otherwise? And why would he attempt to undermine her? It sounds like he is friendly with them, and it's not like he's competing with her in the same market. The finding is a major problem for a huge number of studies...should he just sit on it to keep from hurting her feelings?

The real question is: Who had the idea first? Was it Louise Egan or Dr. Chen? If Dr. Chen had the idea first, and if Louise Egan used his idea to develop her yet-to-be-published experiments, then Dr. Chen has every right to mention it without mentioning her experiments. And this wouldn't discredit her -- since her study is yet to be published -- she could easily publish it and allude to the fact that she's basing her work on the initial problems found by Dr. Chen. Personally, I don't see how someone's work situation is relevant here. It's not.

Hmmm. Lots of anonymous people commenting today. This makes it hard to refer to specific comments.

Re: quite to the contrary.

"It would be scientific misconduct for Chen to make statements based on someone else's unpublished work."

Reasonable people may differ, but I can't agree for several reasons.

1) Chen was speaking via email. I don't know of any codes of conduct about what can and cannot be said by email.

2) Chen's research is also not peer reviewed. So does this rule make Chen's comments about his own work scientific misconduct?

3) Peer review is nice, but it isn't the ultimate judge of quality. After all, Chen's claim is that hundreds of peer reviewed studies are fundamentally flawed. In fact, he's pitting his unreviewed analyses against hundreds of peer reviewed papers, so it would seem that if peer review is the ultimate arbiter, he looses.

I think peer review is very useful, and it is the best system that we have. But it's easy to get hung up on it and distracted from the real issues.

Having just read Chen's complete paper, I have to point out that the analysis doesn't depend on the idea "that the original measurements were in error". Error is a strong word; what's implied is uncertainty in the measurement. No one claims to measure any quantity without uncertainty.

The analysis depends on either of two conditions:
1), that the choices have been given equal ratings. In this case, it's the limited granularity of the rating system that's crucial. To paraphrase the paper, any two items rated '4' are assumed to be equally preferred, but this isn't ever necessarily true. Preferences smaller than the granularity of the measurement scale won't show up if they exist.

2) That people don't perfectly translate their (real) preferences into the preference system being measured. This applies, or may apply, to any different method for measuring preferences. If it is not a completely perfect, uncertainty-free measurement, then some variation between successive measurements would be expected, even without any intervening event such as a choice.

The first condition, the one that's being thrown around on the NYT and other outlets, is not particularly relevant. The second one is much more general, though. The good news is that by making successive measurements in the absence of an intervening choice, the expected variation can be measured statistically for a given preference-ranking method.
Then, if this expected variation is smaller than the effect observed in an experiment with choice, that would be evidence for cognitive dissonance.

The point is that there is a simple and fundamental flaw in many of the papers studying cognitive dissonance, e.g. the Santos m&m paper. The significance of this flaw is that the m&m paper (and others like it) has no interpretable conclusion -- the results could indicate cognitive dissonance or they could indicate an initial asymmetry in the monkey's preferences. The design of the experiment precludes any clean interpretation. Indeed, if the m&m article was submitted for publication now it would not be accepted because no conclusion can be drawn.

Even if there is a recent well designed experiment showing cognitive dissonance does exist, it is still important and worthwhile for Chen to point out the design flaw in these earlier experiments.

["I wouldn't be completely surprised if [cognitive dissonance] exists, but I've never seen in measured correctly." This is hard to believe, since Chen not only works in the same university as Egan, he is a close collaborator of Laurie Santos (Egan's graduate advisor).]

It would be scientific misconduct for Chen to make statements based on someone else's unpublished work. If Santos' unpublished experiments are the only ones that Chen is aware of, then his statement is proper and accurate.

Scientific results don't exist until they have been peer reviewed and published: you don't get credit for them, and others should not refer to them.

(The only exception is if you actually base your work on someone else's unpublished results, but that doesn't seem to be the case here.)

Ye gads- my grant money is on the line! Defend it! Defend it!

Fixed. This is one of those rare moments when Word's grammar check would (might?) have been useful. Thanks.

Please try my web-based experiments

I enjoyed what you've said; however, I'll point out an irksome word choice implying that Socrates is many women. The singular is woman; the plural is women. Socrates couldn't be a women, as you stated.