I would want to know how the sample was collected, and how the questions were posed. This information is very crucial to the validity and interpretation of the data.
Results taken from an ivy league vs. "lower level" university would be expected to differ. It's crucial that the results of this sampling not be extended to the general American (or even state) population. At best, it will be a profile of the type of students which attend that particular university. It's also very important that the sample be collected in a properly randomized fashion - this shouldn't have been too difficult, as the university probably has a list of student names which could be randomly selected. However, if a portion of selected students were unreachable for some reason, you might have a nonresponse bias inherent in the data.
Questions relating to sexual activity seem likely to garner unreliable results, as subjects may lie one way or the other. This is an especially salient concern if there is an actual person going around asking questions - responses to a female vs. a male questioner are likely to differ.
Correlation is a measure of the linear relationship between two qualitative variables, which means that calculating the correlation between things like "Security" and "Conformity" is doubly inappropriate. Not only are neither of these properly quantifiable variables, even if they were, measuring correlation would overlook any possible exponential or logarithmic relationships between them. Before correlation is calculated, the data should be carefully graphed and examined to determine whether a clear linear association is present. I should also repeat the popular "correlation, not causation" warning. Even if a proper negative linear relationship could be established between sexual activity and happiness, this could simply be because of a third, hidden variable. Perhaps a good portion of the students are using some sort of sexual enhancement drugs (say, viagra) which may have adverse effects like depression.
That's all of the stats I can muster right now. Ironically enough, I probably spent more time writing this than I will spend actually doing my stats homework tonight
|
|
Bookmarks