At 03:12 PM 11/26/2008, Scott Woods wrote:
>I want to test a technique involving grammar for its effectiveness 
>in improving reading comprehension.  Please let me know what you 
>think of my design and if you have any suggestions for related research. . . .

DD: You called? From out of the wood work comes there now a 
Psychometrician (Retired.) It sounds like a sound design, as 
presented. However {Ever notice how there is always a caveat?} you 
are dealing with a rather high end sample.* High IQ and all that. Do 
the study. Slap it into a Chi Square contingency test and see. Make 
sure the groups are selected truly randomly. Report the results. I 
personally think you are doing true science here. A lot rarer than 
you would think. The null hypothesis is there will not be a 
significant difference. Fifty fifty. I'd go for the more risky a 
priori prediction that there will be one in favor of the graphic 
syntax. That way I get to use a one tail test and that allows 
significance at a lower level. It is risky, though, because if it 
turns around and bites YOU on the tail and the normal text group 
comes out ahead, the experimenter is required to perform a ritual 
self immolation. We rarely do, though. Just lie and report we did no 
a priori post hoc corrections. Keep me posted. Fascinating to see a 
well designed experiment before it is done. Usually the 
psychometrician just gets a bunch of results and is asked to make 
sense out of it. {Usually by an attractive graduate student. Blonde 
preferred.} I suspect that timing will indeed change the results, but 
that a correlation between speed v extended time will show a high r. 
Suggestion for future reading - anything by Ohmer Milton. I remain, 
your faithful friend and joyous companion in original research.

* Ware the regression toward the mean.

To join or leave this LISTSERV list, please visit the list's web interface at:
     http://listserv.muohio.edu/archives/ateg.html
and select "Join or leave the list"

Visit ATEG's web site at http://ateg.org/