What Situationist Experiments Show

Long time reader; first time poster.   I'm breaking my silence because we've just finished a paper on virtue ethics and the person-situation debate that might be of interest to some X-Philes.  There aren't many philosophers (or psychologists for that matter) who would be interested in a paper about the true meaning of the correlation coefficient and how it relates to the person-situation debate.  But, given the discussion here, I'm guessing that some of you might be.

Here are the details.

What Situationist Experiments Show.

John Campbell, Sarah Meerschaert, Anthony Chemero

Abstract:  Recent critiques of virtue ethics presume that the situationist side is the clear winner in the person-situation debate in psychology. This widely shared presumption is based on some striking findings on the power of situations to influence behavior. However, careful consideration of the empirical evidence and, especially, the nature of correlation coefficients shows that personality traits also account for a large proportion of the variability in human behavior. So the presumed victory for the situationist side is premature at best.

http://edisk.fandm.edu/tony.chemero/campcoughchemsubmit.pdf

Comments most appreciated.

cheers, tony

Comments

[Vishal Lama]

I greatly enjoyed reading your paper that I think presents good arguments that refute the situationist critiques you brought to the casual reader's attention.

The "pencil-picking" example offered in "3. The Nature of Correlations" (p3) illustrated very well the dangers of asking questions that correlation coefficients may not be reliable enough to handle.

Also, arguments offered in sections 5, 6 and 7 are really good ones. Section 5 brings attention to an oft-ignored fact about low correlations sometimes concealing important effects.

Thanks for sharing the paper!

[Ramesh Raghuvanshi]

Man is selfish animal he want to survive in this world in any condition and by any means. Nietzsche rightly told long long ago that all virtues emerge from selfishness.What is motto of selfishness survive in any condition. He may caught in any situation it is dangerous or profitable he behave only how this situation is benefited to him for survival.

[Mark Alfano]

Hi! Interesting paper.

The view attributed to Appiah on page 1 seems to me incorrect. I don't recall his saying any such thing in Experiments in Ethics.

On page two, the following sentence needs to be edited: "After exiting the booth, confederate walked by, ‘accidentally’
dropping sheath of papers." I think it should read "After the participant exited the booth, a confederate walked by, 'accidentally' dropping a sheaf of papers."

Thanks for your informative analysis!

[Alberto]

Hi! Good paper
the point of correlations being related only to rank order has already been made by Vranas on "moral behaviors: cross-situational consistency and predictability"
https://mywebspace.wisc.edu/vranas/web/research.htm

but it seems to me that rank order inconsistency (as proved by low correlation) is compatible with performance consistency (on an absolute scale) only if there is little variation in the level of absolute performance (i.e. everybody has more or less the same level, on an absolute scale), which is not a plausible assumption about traits. If a is 2 on the scale of generosity, b is 8, and c is 15, and they are consistent in their level of performance, how could you possibly have rank order inconsistency? So it seems that in the case of traits rank order inconsistency proves inconsistency in the absolute level of performance.

[Sarah Meerschaert]

Vishal, thanks for the comments!

Ramesh, I haven’t read enough Nietsche but I’ll try to respond to your point. It seems to me that in ethics selfishness is complex. Should I care only for myself? Or selfishly should I act in the interest of those who I care about? Or should my behavioral concerns extend to my family and to my extended family? Obviously this paper does not touch on these issues. My intuitions are that some of our ‘selfishness’ can be seen as evolutionarily based but that our cognitive capacities allow us to move beyond that, but less so in novel situations which are more cognitively taxing.

Mark, you’re very welcome. I think you’re right about the Appiah reference. It’s a mistake that we will be sure to fix. Appiah does pick on these studies somewhat. He says that the psychological claims and data by themselves do not show either that virtues do not exist normatively, or do they explain what's going on in the minority cases (Milgram walk outs and the 10% who didn't let rush stop them from helping in Darley and Baston). And he points out that these studies don't show that dispositions don't exist (49-50). But what we have written makes it sound like he makes a related but different point. Thanks for pointing out that error!

Alberto, I’m having a hard time finding the paper you’ve referred to by Vranas; could you send me a pdf of it? You are right that the difference in rank order does sometimes come along with differences in absolute performance. Our point is just that they don’t always come together. That rank order differences exist doesn’t mean we can infer absolute performance differences (or vice versa).

Thanks for the interest everyone!

[Anonymous]

Hi, interesting paper, although I am not entirely sure whether the more current psychology literature did indeed support the idea of "a substantial case in favor of the situationist side in the person-situaton debate." Your critique focused on articles that were published in the late 60s and early 70s, thus almost half a century ago. A lot has happened since then.

My second point is related to terminology -- it seems to me that using the word "correlation" (a specific statistical test that only applies to measuring two continuous variables, usually with a normal distribution) is not entirely appropriate, since it seems that you are attempting to instead communicate the meaning of "association" (a measure of how any two variables relate to each other). There is a wealth of texts on this on the web if you google for "association vs. correlation."

[Nicole]

Hey great paper, though I admit I could not grasp all of it given my background. I am a statistician and it was interesting to go through this side of correlation coefficient. Will get back with comments, thanks.

[Sarah Meerschaert]

Again, thanks for the comments everyone. As to the recency of the psych literature, I do agree that a lot has happened since the studies we cite. Our point is that these are still commonly cited and influential studies, especially in some of the recent situationist papers in philosophy. We are attempting to cast doubt on the belief that these studies support the situationist argument. While new research is always coming out, as long as these papers are still being discussed their statistics are worth consideration.

I just googled "association vs. correlation" and this link is the third that comes up. Is there a specific book or paper I should look to for more on these terms? I agree that precision in terminology is vitally important in this or any paper. Association is perhaps a better word in the sense that it clarifies when we mean relationships more generally and when we mean a statistical test. But the debate itself is framed in terms of correlation coefficients and this paper may not be the correct place to switch the debate over towards a more exacting jargon.

Nicole, thanks for the consideration and I’d love to hear more about what you, as a statistician think about this whenever you have the time.

[Payday Loans]

We are attempting to cast doubt on the belief that these studies support the situationist argument. While new research is always coming out, as long as these papers are still being discussed their statistics are worth consideration.