In a previous paper (and blog post), I discussed some results (using a new experimental design: “evidence seeking” probes) which support the idea that folk attributions of knowledge are sensitive to stakes (practical interests). This result went against most previous x-phi studies on the subject, but is in accord with a lot of work in epistemology (assuming they make predictions about ordinary people's use of 'knows'). Wesley Buckwalter and others (including Josh May and Jennifer Nagel) made the fair point that my study may possess a confound in that the effect detected is really an effect on belief not knowledge. In a new study, I controlled for belief and continued to find that stakes made a difference to third person knowledge attributions. It is also worth mentioning that this new study aims to control for evidence as well as belief. I have argued here that previous studies may have a problem about controlling for evidence (See also Jason Stanley’s comments here comment #31 and other criticisms by DeRose here). So at least in some respects, I am tempted to think that this sort of study might be more reliable to detect sensitivity than some of the previous work. More details below.
For the original study, participants were given one of two vignettes (Low or High). They concerned a student (Peter) who can proofread one of his English papers before the paper is due. In the High condition, unusual circumstances make it so that if he has a single typo, he will lose his scholarship and this is very bad for him. In the Low condition nothing bad happens but he would still prefer to have no typos. Participants were then asked “How many times do you think Peter has to proofread his paper before he knows that there are no typos? ____ times”. It turns out that participants gave median answers of 2 and 5 for the Low and High conditions respectively. The difference was highly statistically significant. I took this to support the idea that folk attributions of knowledge are sensitive to stakes (this study was also replicated by Buckwalter). Now Buckwalter and others objected that the effect detected was really about belief not knowledge. The idea is that "high stakes" Peter will collect more evidence before settling on belief (a phenomenon well studied by psychologists and famously brought to bear on these issues by Jennifer Nagel).
As a response to this suggestion, I ran a further study that controls for belief. The vignettes were the same but I asked the following question (after asking some comprehension questions): “It turns out that right after Peter finished writing his paper, he formed the belief that there are no typos in his paper—and in fact there are no typos. But does he know this? How many times do you think that Peter has to proofread his paper before he knows there are no typos? ___ times.” Here again, people tended to give lower answers in the Low case than in the High case. And this difference is highly statistically significant Mann-Whitney (N=78) U=472.500 p=.003 (Mean Low= 2.44, Median Low= 2, Mean High=3.62, Median High=3). In my view, this result is explained nicely by the idea that folk third person attributions of knowledge are sensitive to stakes in the way that IRI would predict if the folk also accepted (or behaved as if they accepted) some principle connecting knowledge and action. I should also add that there is a stat sig. difference between my original high stakes case and the new high stakes case. This might very well be because Buckwalter and others were partially correct in that *some* participants were focusing on belief not knowledge. But an alternative explanation is that in the new study I am reporting today (but not in the old one I ran) participants were told that when Peter finished the paper there were no typos. I think this gives some reason to think that participants will think Peter will need to proofread less times (than in the original high stakes). So what do you guys think of this new evidence (going against the tide in x-phi) that folk attributions of knowledge (and not just "belief") are sensitive to stakes?
No comments:
Post a Comment