I have a question/request that I know is unusual but that I'm hoping you'll consider. I help to lead a small group of researchers at the Minnesota Center for Twin and Family Research at the University of Minnesota interested in understanding effects of substance use and abuse on brain development in adolescence and early adulthood. Alcohol consumption can be characterized in different ways and it isn't immediately apparent what is most relevant for effects on brain structure and function. We have therefore created an index of alcohol use that combines 4 different aspects of exposure: typical quantity consumed in the past year, frequency of past-year consumption, density of consumption (maximum number of drinks consumed in a single 24-hr period), and misuse (number of times drinking to the point of intoxication). We have "ordinalized" the distribution of responses to these 4 different questions by grouping together similar responses, and the index derived from these ordinalized versions of the original variables has good psychometric properties and appears to work well as a measure of alcohol exposure; we are obtaining several interesting associations in our sample of twins with MRI and EEG data, including apparent causal effects of exposure in twin-difference analyses.
I personally am interested in using data from the HCP, and I'm part of a different group here at the U of MN doing so with approval from the IRB here. I'd like to create a similar index of alcohol use from the relevant SSAGA questions asked of HCP participants. However, responses to these questions have been categorized/ordinalized differently than we have done at the MCTFR. In particular, the HCP version has greater sensitivity at the low end of alcohol use, whereas the MCTFR version has greater sensitivity at higher levels of use. I'm interested in replicating findings in one data set in the other data set, or in combining results of the two analyses through meta-analysis. To do so, it would clearly be best to use as similar a method of categorizing responses in both as possible. Given that we have good evidence of the effectiveness of our approach, it would be ideal to use a similar method with the HCP data. However, that would require having the original, uncategorized variables in order to create categories similar to ours. I know that this is an unusual request, but is it possible to obtain a file with the 3 or 4 original variables I'm interested in? I'd obviously be happy to answer any questions or address any concerns you might have about this. Sincerely, Steve ----- Steve Malone Research Assistant Professor MN Center for Twin & Family Research Department of Psychology University of Minnesota 612/626-1436 (voice) 612/626-2079 (fax) _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
