CHICAGO (Reuters) - Male children born to women who smoke during pregnancy run a risk of violent and criminal behavior that lasts well into adulthood, a study published Sunday said. Their study -- based on a look at the arrest histories up to age 34 of 4,169 males born between 1959 and 1961 in Copenhagen, Denmark -- was the first to show that the impact lasted beyond adolescence into adulthood.
Correlation studies are one of the most fascinating fields in statistics from the results they find. It is also the most misleading. As I stated in the Glossary under "Correlation Cause and Effect", there are three possible reasons for the correlation: 1. A causes B, 2. B causes A, and 3. Some unknown factor C causes both. In the above study we can easily rule out number 2, since violent adults cannot cause their parents to smoke during pregnancy. So the correlation has to be the result of 1 or 3. The makers of the study had their own conclusions:
The study said the mechanism behind the effect might be damage done by smoking to the central nervous system of the fetus. The effect uncovered in the study persists even after accounting for such factors as socioeconomic status, parental psychiatric problems, age and the father's criminal history.
Thus, the study favors possibility 1 over 3. Still, there are other possibilities that cannot be ruled out. Maybe females that smoke when they are pregnant are more likely to be bad mothers. One of the things that make correlation studies so popular, is that the causation is vague. Therefore, you can use causation to prove any political position you want.
Correlation studies are also easy, anyone can pick up an Almanac and create one. Just grab two seemingly independent statistics broken down in a similar way (like geography, age group, year, etc.), then create a scatter graph plotting the statistics against each other. If the scattered dots look like this [/], then you have a positive correlation, if they look like this [\], you have a negative correlation. If they look like this [-] or this [#], then they do not correlate at all. This is more or less what the study above did.
"Compared with males whose mothers did not smoke during the third trimester, males whose mothers smoked more than 20 cigarettes (a day) during the third trimester were ... 1.6 times as likely to be arrested for nonviolent crime ... 2.0 times as likely to be arrested for violent crime and ... 1.8 times as likely to be life-course persistent offenders,'' the researchers found. They noted ``a linear relationship ... between percentage of violent offenders and number of cigarettes the mother smoked.''
A few years ago, an article I read named the the states with the highest and lowest crime rates. The states with the highest: Florida, Texas, Arizona, California, and Georgia. The states with the lowest: West Virginia, North Dakota, South Dakota, Kentucky, and Pennsylvania. What the highest crime states all have in common is also high growth rates. The exact opposite is true with low crime states. So, I decided to do my own correlation study. I used 1990 figures, because they were the most recent year that I have both growth and crime statistics. There was indeed a positive statistical correlation of 0.44. Click here for the data used. Note that with only 50 data points, this is not entirely scientifically valid.
The question is why do states with high growth rates also have high crime rates? I believe it has to do with when people move away from family and friends, they are more likely to lose control. The divorce rate also correlates to growth rate.