I agree that they are separate concepts however they are not completely independent. There are times when it makes perfect sense to combine a number of discrete measurements into continuous data. An example would be to count the number of correct responses on a multiple choice exam. While each response was discrete (i.e right or wrong) it makes sense to count them up & report an "average" number of correct responses. However, it makes to sense to look at the number of people who chose A, B, C, & D for a particular question and then report the "average" response to the question was "B.2". In the Holland study I feel that the data are more like the latter scenario. Had there been repeated measures I might be more willing to accept that Ss sat 2.2 seats away, but with only one trial I still think that a non parametric analysis would have been better.
-Don. Christopher Green said: > It seems to me that a couple of people at least are confusing "levels of > measurement" with the question of discrete/continuous data. They are > quite distinct concepts. It is perfectly possible to have a discrete but > interval (or even ratio) scale. Number of chairs (per se) would, in > fact, seem to be a fine example. If you doubt this, consider if you > wanted to measure, say, the popularity of a various performer by > measuring the number of people who attended their concert (where the > typical value would be in thousands, even tens of thousands, rather than > just 1-4). Wouldn't it be reasonable to say that (other things being > equal) a performer who draws 10,000 people is twice as popular as one > who draws 5,000 people? If you think yes, then you believe in discrete > ratio variables. If you are tempted to say no on account of some > presumed methodlogical flaw, then you probably didn't understand the > meaning of my parenthetical insertion of "other things being equal." If > you say no on the basis of some other (non-methodological) grounds, then > I'd be interested to hear them. > Regards, > Christopher Green > ==================Shearon, Tim wrote: > >>Hmmm- We all know what zero chairs would be but does this meet the >> requirement of a true interval scale- barely. Integer differences which >> are translated into not possible interpolations (what is .387 chair >> corresponding to?). That's a problem. EXTREME restriction of range (not >> just restriction) is a real problem. Independence? I see lots of >> problems with the assumptions of ANOVA on this one. I think the use of >> non-parametric statistics is required. Perhaps some areas use the "opt >> outs" of Monte Carlo studies and robustness of the ANOVA etc. to get >> beyond reason here. But remember these are usually relying on the >> presence of large numbers whereas this study seems to rely on rather >> small ones. These small numbers are also, as Stephen and others have >> pointed out, include some VERY plausible selective attrition from the >> experimental but not the (oddly) combined control conditions. All in >> all this should not have been reported in this form. As others have >> shown there IS NO STATISTICALLY SIGNIFICANT effect there unless you >> blatantly manipulate these two control conditions into one. As a >> reviewer I could never approve of such and I'd certainly return any >> paper to a student an require the correction of the statistics. Maybe >> this could pass muster as a "thoughtful" early result but this is ONE >> study coming after a great deal of research and a large number of >> studies trying to find support for this effect. And as others have >> pointed out such results are highly likely to be transient. Perhaps, >> and I am really reaching here, there is enough suggestion to continue >> for long term study of the effects of extended education. But I'd >> certainly never approve funding for anything based on such >> manipulations and over-cooking of a very small difference. (Anyone gone >> to the trouble to look at the effect size here- it is easy to compute >> and quite informative. Why didn't any of the reviewers go to the >> trouble to do that I wonder- Is it that they already believe in the >> effect? Makes me wonder anyway.). :) Tim Shearon >> >> -----Original Message----- >> From: Wallace Dixon [mailto:[EMAIL PROTECTED] >> Sent: Mon 8/23/2004 11:26 AM >> To: Teaching in the Psychological Sciences >> Cc: >> Subject: Re: First Solid Evidence that the Study of Music Promotes >> Intellectual Development >> >> >> >> Don, >> You lost me on this one. How is it NOT interval, even ratio, >> data? I >> can see it would be easy enough to have qualms about restriction in >> range, etc., or even qualms about using NHST at all, but I don't get >> how distance in "seats away" isn't ratio? Sorry for being so dense. >> >> Wally Dixon >> >> >> >> >> On 8/23/04 1:12 PM, "Don Allen" <[EMAIL PROTECTED]> wrote: >> >> > I quite agree, and I wish it were the only lapse in APS editing. In >> Holland et al. "Don't Stand So Close to Me: The Effects of >> Self-Construal on Interpersonal Closeness" ( Psychological Science >> Volume 15 Issue 4 Page 237 - April 2004 ). They report the >> following methodology: >> > >> > "After completing the lexical decision task, the participants were >> asked to take a seat in a waiting area, ostensibly to give the >> experimenter some time to prepare the second part of the experiment. >> Four chairs were lined up in the waiting area, with a jacket hanging >> over the chair on the extreme left. This jacket suggested the >> presence of another person (Macrae & Johnston, 1998). The dependent >> variable was the distance, *** in number of chairs, *** (my >> emphasis) between the chair with the jacket on it and the chair that >> the participant chose to sit on." >> > >> > They then analyse the data as follows: >> > >> > "To examine the effects of self-construal and gender, we performed a >> 2 (self-construal: independent vs. control)2 (gender: female vs. >> male) between-subjects analysis of variance on the distance between >> the participant's chair and the occupied chair. As expected, >> participants in the independent-self condition sat further away >> (M=2.07) than participants in the control condition (M=1.66), F(1, >> 73)=8.57, p<.01. 1 No main effect of gender was obtained. Also, no >> interaction effect was found." >> > >> > Now I have a hard time accepting that "number of chairs" is interval >> data. A non parametric analysis would have been far more >> appropriate. Editorial rigour just ain't what it used to be. >> > >> > -Don. >> > >> > >> > >> > Stephen Black said: >> >>> Ronald C. Blue wrote: >> >>> >> >>>> http://www.psychologicalscience.org/media/releases/2004/pr040819.cfm >> >>> First Solid Evidence that the Study of Music Promotes Intellectual >> Development >> >> >> >> and Chris Greeen commented: >> >> >> >>> Now none of this is out and out "wrong," >> >> >> >> Oh, it's wrong all right. See earlier exchanges on this "solid >> evidence", between Ken Steele and me, for example, at >> >> http://www.mail-archive.com/[EMAIL PROTECTED]/msg10749.html >> >> >> >> Stephen >> >> >> >> ___________________________________________________ >> >> Stephen L. Black, Ph.D. tel: (819) 822-9600 ext 2470 >> Department of Psychology fax: (819) 822-9661 >> >> Bishop's University e-mail: [EMAIL PROTECTED] >> >> Lennoxville, QC J1M 1Z7 >> >> Canada >> >> >> >> Dept web page at http://www.ubishops.ca/ccc/div/soc/psy >> >> TIPS discussion list for psychology teachers at >> >> http://faculty.frostburg.edu/psyc/southerly/tips/index.htm >> >> _______________________________________________ >> >> >> >> >> >> --- >> >> You are currently subscribed to tips as: [EMAIL PROTECTED] To >> unsubscribe send a blank email to >> >> [EMAIL PROTECTED] >> > >> >> ---------------------------------------------------------------------------- >> Wallace E. Dixon, Jr. | >> Chair and Associate Professor | Understanding atomic physics >> is >> of Psychology | child's play, compared with >> Department of Psychology | understanding child's play. >> East Tennessee State University | -Albert Einstein >> Johnson City, TN 36714 | >> (423) 439-6656 | >> ---------------------------------------------------------------------------- >> >> >> >> --- >> 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] -- --- You are currently subscribed to tips as: [EMAIL PROTECTED] To unsubscribe send a blank email to [EMAIL PROTECTED]
