Thursday, February 11, 2016

More Fun With Stats

I mentioned a few days ago my discovery that burglary and murder rates in the U.S. for 1960-2012 are strongly correlated (r=0.934).  Now I did a little fiddling with Excel, and I find the ratio of burglary rates to murder rates during that period ranges from 99.72549 to 168.0816.  Mean is 138.7043, and standard cdeviation is 16.19641.  I also believe I have identified why this relationship exists.  Much of U.S. murder is drug-trafficking related: gang members killing each other in disputes over territory; much of burglary is committed by people with addiction problems they are trying to pay for.  In short, this relationship showing social disorder has a common cause: the level of illegal drug addiction in a society.

I really hate  what they did to Excel's charting tool a while back.  I want to use three columns of data: year, log10(burglary rate), log10(murder rate) and produce a chart with year across the bottom and the two log value series as lines.

Figured it out.  Still clumsy.


No comments: