Coyote Blog

A while back, I pointed out that immigration opponents seemed to be depending on American's having poor match skills and a pathetic knowledge of history.  Today in this post from Captain's Quarters we find more statistical funny business.  Captain Ed, like many conservatives, have been stumping for the US to build a big honking fence at the border, nominally as part of the war on terrorism.

Of course according to supporters it is only about security, not xenophobia, which explains why the fence proposal in Congress covers both our northern and southern borders since both are equally porous to terrorists.  Oh, wait, the law only covers the southern border?  Oh.  Well, I hope terrorists can't read a map and don't notice that the northern border is three times as long and in many cases more unpopulated and unguarded than the southern border.

Anyway, another "security" argument by immigration foes is that hordes of criminals are apparently pouring across the border, and walls are proposed as a way to stop them.  The Captain quotes Bill Frist:

One of the most important and most effective ways that we can stop
illegal immigration is through the construction and proper maintenance
of physical fences along the highest trafficked, most commonly violated
sections of our border with Mexico.

Take the case of San Diego. According to the FBI Crime Index, crime
in San Diego County dropped 56.3% between 1989 and 2000, after a fence
stretching from the Ocean to the mountains near San Diego was
substantially completed. And, according to numbers provided by the San
Diego Sector Border Patrol in February 2004, apprehensions decreased
from 531,689 in 1993 to 111,515 in 2003.

Whoa. That sounds impressive.  But, remember what I often say on this site -- correlation is not causation.  Indeed, it is not just random chance that he picked the years 1989 - 2000.  Those were the years that nearly every part of the US saw a huge drop in its crime rate.  Using this data for these years, and presuming Frist is using the crime rate index per 100,000 people, which is the stat that makes the most sense, here are some figures for 1989 - 2000:

Crime Rate Change, 1989-2000:
US :  - 28%
Arizona:  -28%
California: - 45%
New York: -51%

Wow!  The border fence in San Diego even had a similarly large effect on crime in New York State!  That thing is amazing.  Oh, and note these are state figures.  My understanding is that the figures for large metropolitan areas is even more dramatic.  So what happened in 1989 to 2000 is every state and in particular every large metropolitan area in the country saw huge double digit drops in crime, and San Diego was no exception.   But Frist tries to give credit to the border fence.

In case you want to believe that Frist does not know what he is doing with these stats (ie that he wasn't intentionally trying to give credit for a national demographic trend to a border fence in San Diego) notice that 1989 was the US crime rate peak and 2000 was the US crime rate low point.  So with data for the years up to 2005 available, he just happens to end his period at 2000.  Oh, and the new style fences he wants to emulate were actually only started in 1996 (and here, search for "triple fence"), AFTER most of these crime gains had been made.  Correlation definitely does not equal causation when the proposed cause occurred after the effect.

For all of you who always wanted to live in Soviet East Berlin, you may soon get a good taste of that experience:

The first fence, 10 feet high, is made of welded metal panels. The
second fence, 15 feet high, consists of steel mesh, and the top is
angled inward to make it harder to climb over. Finally, in high-traffic
areas, there's also a smaller chain-link fence. In between the two main
fences is 150 feet of "no man's land," an area that the Border Patrol
sweeps with flood lights and trucks, and soon, surveillance cameras.

Below are views of Nogales, AZ and Berlin.  Nothing alike.  Nope.  Totally different.

Finally, I will give the last word to Frist, bold added.

That's why I strongly support the Secure Fence Act of 2006 "¦ and that's
why I'm bringing this crucial legislation to the floor of the Senate
this week for an up-or-down vote. By authorizing the construction of
over 700 miles of two-layered reinforced fencing along our southwest
border and by mandating the use of cameras, ground sensors, UAVs and
other forms of hi-tech surveillance, this legislation would help us
gain control over every inch of our borders "“ once and for all.

"gain control over every inch of our borders," except, or course, for those 3000 5525 miles (350 million inches) to the north where the people on the other side have the courtesy not to speak a foreign language.  But its hard to demagogue well about a threat from Canadians, since they are mostly WASPs like we mostly are, or at least it has been for the last 100 years or so.  54-40 or fight!

Update: Here is that terrifying Canadian border barrier (from this site).  This demonstrates why our terrorist security dollars need to all be invested on the southern border, since this one is already locked down tight.  Heck, there is one of these babies (below) every mile!  Beware terrorists!

And don't forget these terrorist-proof border checkpoints along our northern frontier:

But it's not about race.

Update 2:  Yes, my emailers are correct.  I did not actually give Frist the last word like I said I would.  Gosh, I feel so bad about that.

Update 3:  Welcome to readers of my favorite site, Reason's Hit and Run.  It looks like Texas may soon consider a border fence, though with Louisiana instead of Mexico.

I didn't really pay much attention to the typical 24-hour partisan finger-pointing flurry accompanying the stats showing an uptick in US infant mortality rates.  However, looking back on it, it is a good lesson about how statistics are routinely misused in this country (via Captain's Quarters).  Critics of the administration and the health care system used the statistics to try to show something is wrong with the US. 

Two weeks ago, Nicholas Kristof wrote a column on the first increase in the American infant-mortality rate in decades, taking the opportunity to excoriate Americans and the Bush administration as uncaring and unresponsive to the deaths of children. He compared the US unfavorably with Cuba and China

Unfortunately, this conclusion was flawed:

babies that would die in the womb or at stillbirth elsewhere are born alive in the US. Many of these survive completely, but because of their precarious state, they tend to die at higher percentages than normal births. That's why the numbers rose slightly for 2002. The CDC doesn't expect to see another increase like it.

The solution might be to look at the survival rate at year one as a percentage of total pregnancies, not just births (though you would have to exclude abortions).