Hi I think this paradigm is susceptible to a statistical artifact that I have noticed in some other parapsychological experiments (Schwartz's work comes to mind). The researchers essentially select distinct conditions, defined in part by subjects' responses (an important point!), and then determine whether the observed hit rate in that condition differs from 50%. But the chance value may not be 50% if subjects demonstrate any consistent preferences in making their responses.
I can illustrate with hypothetical data for the paradigm in the linked document. If I'm understanding correctly, the paradigm is essentially that subjects indicate a preference for one of two pictures, and then one of two pictures is randomly chosen for subliminal presentation. For certain types of trial (e.g., unpleasant items), the preferred item has a probability greater than .5 of being the item later selected for subliminal presentation. But .5 is only the proper comparison if subjects have an equal probability of picking the different types of pictures. If they have a preference for one type of picture over the other types, then the chance probability of a match is greater than .5. Suppose there were 200 trials in an experiment, with 100 negative and 100 positive pictures. And suppose that on half of the negative trials (50 of 100), the negative picture is selected for subsequent presentation. If subjects choose negative stimuli 50% of the time, then the number of expected matches is .5 x 50 = 25, which gives an expected probability for this cell of 25/50 = 50%. But if subjects choose the negative stimulus more than 50% of the time, then the expected probability changes. For example, if subjects choose the negative stimulus 60% of the time, then the expected number of matches is .6 x 50 = 30, which gives an expected probability for this cell of 30/50 = 60%, not the 50% against which researchers are wrongly comparing the observed proportion of hits. I could not find anything in this article to indicate what the baseline choices of the different kinds of stimuli were, so this remains hypothetical. But one should be extremely cautious about .5 being the chance level anytime outcomes are defined partly by participants' responses. It is unfortunate that someone of Bem's stature is giving credence to this nonsense. It is scant comfort to remember that Newton also believed in astrology (I believe), inasmuch as astrology may now be a thing of the past if he had not (unlikely, of course, but one can only hope!). Jim James M. Clark Professor of Psychology 204-786-9757 204-774-4134 Fax [EMAIL PROTECTED] >>> [EMAIL PROTECTED] 14-Nov-05 3:40:09 PM >>> A leading psychologist publishing in at least one leading journal (Psych Bull), offering his program to anyone interested. I can't say I'm convinced, but he's pretty darn convincing. Now, I'm off to buy a lottery ticket. Tomorrow after the numbers are announced, I'm going to present them to myself subliminally. ________________________________ From: Rick Stevens [mailto:[EMAIL PROTECTED] Sent: Mon 11/14/2005 4:32 PM To: Teaching in the Psychological Sciences Subject: Re: student's question Cancel that suggestion. I just put 'precognitive habituation' into Google and it is Daryl Bem that I found first. http://www.dbem.ws/Precognitive%20Habituation.pdf I haven't read the whole thing but it says that anyone with a computer and a stat program can reproduce it. I'll be a millionaire in no time. Rick Stevens wrote: Burns, Daniel wrote: Research suggests that subliminal perception does influence behavior (e.g., mere exposure effect and sublimal/masked priming effects), as well as emotion (e.g., mere exposure to stimuli increases one's mood). In fact, recent research out of Cornell suggests that stimuli presented in the future actually affect your behavior now (e.g., subliminal precognition). I think that we have a candidate for the James Randi million dollar prize. Send those people to Randi.org and have them sign up. -- __ Dr. Rick Stevens __ Psychology Department __ University of Louisiana @ Monroe __ [EMAIL PROTECTED] --- You are currently subscribed to tips as: [EMAIL PROTECTED] To unsubscribe send a blank email to [EMAIL PROTECTED] --- You are currently subscribed to tips as: [EMAIL PROTECTED] To unsubscribe send a blank email to [EMAIL PROTECTED] --- You are currently subscribed to tips as: [email protected] To unsubscribe send a blank email to [EMAIL PROTECTED]
