If each of the group scores were converted to z-scores, wouldn't the z-score
mean for each group (by definition) be equal to zero? So you would end-up
comparing two zero scores (with a standard deviation of 1.00 for each group).
Does your student want to know whether participants performed similarly on each test? If that's the case, might a correlation be performed across the two tests (then absolute scores and chance levels would not matter). "Manza, Louis" wrote:
--
"unanswered questions are less dangerous than unquestioned answers"
|
- correcting for chance Patrick O. Dolan
- RE: correcting for chance Manza, Louis
- RE: correcting for chance Steven Specht
- RE: correcting for chance Peterson, Douglas
- Re: correcting for chance Patrick O. Dolan
- Re: correcting for chance Donald McBurney
- RE: correcting for chance Manza, Louis
- Re: Correcting for chance Herb Coleman
- RE: Correcting for chance John Kulig