Monday, January 25, 2010

Correlation Does not Equal Causation - Lancaster, Take Heed

Too often, we fall prey to the fallible argument - as intuitive as it may be, at times - that when two things correlate with each other, by necessity one of those things causes the other to occur. A made-up example: In cities with more than 60 street lamps per 1,000 residents, the rate of heart disease is 2.5 times higher than in cities with less than 60 street lamps per 1,000 citizens. You would not make an argument that street lamps, in of themselves, cause heart disease. At least not seriously, one would hope.

Being human, our elected officials sometimes argue fallaciously about correlated events. In Lancaster, California, Mayor Rex Parris has most recently suggested that the mandatory castration of Pit Bull and Rottweiler type dogs caused a reduction in crime, most notably gang-related crime in 2009. The Mayor cites the Los Angeles County Sheriff Department's statistical data, but 2009 information is currently unavailable to the public. You can see previous years here.

Where are the numbers coming from?
Mayor Parris states that "Lancaster’s violent gang crime, which includes homicide, rape, robbery and aggravated assault, fell by 45% last year, and there was a drop in overall gang crime by 41%". It is difficult to challenge this without the data in front of me. According to the city's press release in October of 2009, the crime rate had declined 23% in the first nine months of 2009. While not impossible, it does seem unlikely that, in the last three months of 2009, Lancaster was able to boost their reduction in the crime rate from 23% to an amazing 45%. The Los Angeles County Sheriff's statistics available to the public do not delineate between "gang-related" crimes and all crimes. Without more specific numbers, it is difficult to garner much information from Parris' statement, except that gang-related crimes in 2009 were reduced. I can only assume Parris and other mayors are privied to other statistical information that assists in their "gang-related crime"assertions.

Crime reduction already on its way:
Lancaster's number of severe crimes* increased dramatically in 2003. In 2002, 6,961 serious crimes were committed in Lancaster and in 2003, that number grew to 8,301. In 2007, the numbers peaked and, in 2008, Lancaster saw its first significant reduction in the number of severe crimes since 2002. There were 685 fewer serious crimes in 2008 than in 2007. More importantly, the crime rate (that is, the number of crimes committed per 10,000 people) started to decrease in 2007. And continued in 2008.

Another way of putting it: A full two years before Pit Bull and Rottweiler type dogs were required to be castrated, Lancaster was seeing a reduction in its crime rate. And in 2008, Lancaster was seeing a reduction in the number of serious crimes committed, including gang-related crime.

Certainly, Mayor Parris would not suggest the mandatory sterilization bill worked retroactively in reducing crime.

Crime rate reduction occurring elsewhere:
Palmdale is located in the Antelope Valley, along with Lancaster. They have approximately the same population sizes. Their demographics are similar in some respects - they have between a $41-45,000 median income. Their median income for a family is nearly identical. Lancaster does have a higher percentage of people identifying themselves as white than Palmdale. Both have the same number of single moms. In many ways, Palmdale and Lancaster are quite similar. But in their crime rate, they are different. Palmdale has had a consistently lower crime rate than Lancaster.

More importantly, during the same time that Lancaster's crime rate started to decline...well, so did Palmdale's. According to their press release (and again, I do not have access to "gang-related" crime #s), there was an overall 25% reduction in gang-related crimes, a 67% reduction in homicides and a 41% reduction in gang-related robberies.

Palmdale does not have any laws regarding the mandatory sterilization based on breed.

In fact, in 2008, the entire Antelope Valley saw a 10% reduction in the overall crime rate - the greatest drop since 2001.

Who do these dogs belong to?
In 2009, as a result of the ordinance, 1,138 Pit Bulls and Rottweilers were impounded by the Los Angeles County Department of Animal Care and Control on behalf of the City. Of those, 362 were voluntarily surrendered by their owners in response to the strict City ordinance.
Mayor Parris takes these numbers as evidence enough that this law reduces gang-related crime. Parris assumes that gang-members would voluntarily surrender their dogs to the shelter. He assumes gang-members care about being caught with an intact Pit Bull or Rottweiler. And he assumes that a seemingly large percentage of the "gang-banging" population will just stop their habitual criminality and start becoming upstanding citizens because they don't have intact Pit Bulls or Rottweilers. The failure of logic is astounding.

I do not have the individual stories in front of me, but I will hazard a guess at who exactly is dumping those 326 Pit Bulls and Rottweilers - some certainly may be gang members. Most are probably people who are a) confused by the law; b) incapable of affording or unwilling to fork out the cost of castration or c) disinterested in owning a castrated non-puppy-producing (and thus money-making) dog. I could be going out on a limb here, but my guess is that most of these people are not gang members.

Even if all 326 owners were gang members, there is still no statistical evidence that their dogs being placed in the shelter and a) adopted or b) killed reduces gang-related crimes. None. Raw data are what they are and can certainly be construed however the reader sees fit. Data that have been analyzed? With 95-99% confidence intervals? With the elimination of all other variables that might confound the data? That is scientific evidence worthy of writing news articles about.

But this? This haphazard, illogical, unscientific, non-statistical manipulation of raw numbers? It's a great way to set people and animals up for failure. The entire premise results in the death of dogs and the criminalization of certain dog guardians/owners. It creates a false safety net wherein other council members or mayors may try to improve safety by implementing an unrelated, uncorrelated law. All without any factual evidence that castrating Pit Bulls and Rottweilers causes a reduction in gang-related crime. Unless Palmdale residents thought they too had to castrate their Pit Bulls and Rottweilers, it seems - to me, anyways - that the cause of gang-related crime rate reduction had little to do with the testicles and ovaries of dogs.

What is crime?
It should be noted that throughout 2009, Mayor Parris and company seem to suggest that there were other reasons for the reduction. Not once do they mention doggy breeding body parts. They mention a lot of the new programs the LA County Sheriff's Department has implemented in the past two years.

We know crime is not a one-variable event. Fixing broken windows, in of itself, will not eliminate crime.  Increasing the number of police officers. Reducing the time it takes to reach the scene of a crime. Increasing social welfare programs. Improving the economic value of impoverished regions of cities. Inviting and encouraging civic participation with programs like the Neighborhood Watch. Reducing the likelihood of juvenile mischief through alternative after-school programs. Increasing graduation rates. We know that the likelihood of committing serious crimes depends on a host of variables - where you were born, who you were born to, your education, your economic background, your friends and social peers, the range of options you have after school (both after high school and, literally, when the school day ends), etc. ad naseum.

A complex social phenomenon, like crime, cannot be reduced to something so simplistic as whether Pit Bulls and Rottweilers have testicles or ovaries. It is not only illogical, it's offensive to our collective intelligence.

*Which the LA County Sheriff's Dept defines as Part 1 crimes include homicide, forcible rape, aggravated assault, grand theft auto, burglary, larceny and arson.


kcdogblog said...

This just seems like a case where they want to justify the law so badly that they're willing to believe almost anything to make it ok.

The whole notion is absurd. I can't even get my head around the idea that someone who would commit a murder, rape, or other violent crime would move from a city because of a misdemeanor law that required their dog to be altered. If they aren't obeying murder laws, chances are they're goign to take their chances on the dog one as well -- not to mention, you should be using SWAT teams to deal with gang activity -- not animal control officers.

Anyway, it's absurd.

Rinalia said...

It is absurd and offensive that the mayor is trying to pawn off his silly law as the ANSWER to gang-related crime. It goes beyond logic.

Pibble said...

It's almost as illogical as BSL. If the Lancaster mayor believes that mandatory castration laws reduced crime, he would also believe that BSL would reduce the number of [enter breed here] in his community, therefore reducing [enter breed here]-related attacks.

Instead of addressing the true problem – irresponsible owners or criminals - he’s incredibly ignorant and does a huge disservice to his constituents.

Law abiding citizens obey the laws put into place, but they're not the problem, are they? Criminals, thieves, thugs - whatever you want to call them - look at these laws as a challenge, and exist to smash them in the faces of people like this.

PoochesForPeace said...

I love your knowledge of research! This is such a simple yet important rule in the world of research.

Rinalia said...

@Pibble: I'd have to say it's even more irrational than BSL! I mean, at the very least, the premise of BSL is to target biting dogs and address the issue of dog bites (which it does poorly, of course). Parris' argument is that a law having to do with dog testicles magically reduces homicide, rape, aggravated assault and larceny!

@PoochesforPeace: My college statistics professor hounded us about correlation and causation. HOUNDED! The lesson has stuck with me!

Pibble said...

Agreed! My basis for comparison was weak, a grasp at straws, even. :p

But Mr. Mayor is still a sad excuse for a community leader. When I think of the sad lambs that might believe his statements, the little hairs on the back of my neck stand up.