Talk:Correlation/Introductory quiz

Guns - Crime relationship

The answer to question 3 is incorrect, and I do not know how to edit the quiz to fix it. Data indicates that there is a slightly negative correlation between gun ownership and gun deaths. See "More Guns = More Crime?"[1] for pointers to data on this topic. (Note -- most responsible gun owners know this.)

Due to the political nature of this question, maybe question 3 should be changed to one that most people can estimate correctly.--Target4cactus 01:51, 16 September 2009 (UTC)

<<>>

I have a MS in statistics and I am working on a Phd in statistics at a major university. In my graduate career I have received one B+ and everything else As; and I missed problems 3 and 4 =) I put small positive and small positive on both. I think you arn't testing statistics knowledge there.

I am actually very suspicious of your claim that gun deaths are strongly correlated with guns in a community. Actually, now that I got warmed up I think you are smoking dope to claim gun deaths have a strong linear correlated with gun counts in a community. I mean think about it. Why would that be a tight association? Imagining a pairs plot of guns in a community versus gun deaths in that community I would expect noise out the fricken wazzu. That my friend is not a good example of moderate to high correlation =)~ Your robberies example suffers the same way when you get "real" instead of "textbooky". Your talking a social science phenomenon. If you actually were to plot robberies versus incidences of drug abuse with real community data you arn't going to see anything remotely tight there either. Even "moderate" would be way to strong a word. Social sciences have notoriously low R-Squared values.

Agreed - thanks. I've changed

• #3 to "What would you expect the correlation to be between daily calorie consumption and body weight to be?" (moderate to large positive)
• #4 to "Estimate the correlation between driving performance and blood alcohol levels" (moderate to large positive)
• #6 to "Estimate the correlation between psychological well-being and psychological distress" (moderate to large negative)

What do you think? Appreciate your thoughts. -- Jtneill - Talk - c 06:01, 6 March 2010 (UTC)