 |
 |
I have a hypothesis in my field of neuropsychology. I think that a certain experimental treatment might be increasing the variance of scores compared to a control condition, but not moving the means all that much, which is why many investigators are coming up with null results for the treatment. There are nice a priori theoretical grounds why this might be so, and this is not some post hoc attempt to explain some awkward results. I am imagining that a particular remembered stimulus might "push" or "pull" on a subsequent response, depending on whether it is remembered in the correct or incorrect context - hence if it is remembered at all, it will affect subsequent response time either positively or negatively, whereas if it is forgotten, it will have no effect one way or the other. How can I go about testing this hypothesis? Many people in my field have raw data that could be reanalysed from this point of view, since they were all looking for mean differences and not effects on the variances. Anything not overly sophisticated that folks could suggest to me? TIA, David Klein %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%