Kicking Ass (aka Stop and Frisk) – Deterrence or Labeling?

August 2, 2013
Posted by Jay Livingston

Critics of stop-and-frisk claim that the policy, when used on a large scale, is counter-productive. Being stopped and frisked is not a pleasant experience, and the vast majority of people searched are not carrying illegal weapons or drugs.* To them, it just looks as though the police are “throwing their weight around.” 

The critics argue further that these aggressive police tactics reduce the cops’ effectiveness in doing what cops are supposed to do –  catch criminals and prevent crime.  For that, the police and the city need the help of ordinary people. If the community is largely alienated from the police and the government they represent, people will be less likely to help the police. 
                                                       
The counter argument is that stopping a large number of people in the pool of potential criminals – i.e., young males – will reduce crime not only among the tiny fraction that are arrested but among the others as well.  Police weight-throwing will act as a general deterrent. As the cop says (the one approvingly quoted by Wilson and Kelling, in their classic “Broken Windows” essay), “We kick ass.”

Does kicking ass deter, or does it alienate?  It would be nice to have evidence rather than assertions.  A recent study by Stephanie Wiley and Finn-Age Esbensen speaks to this very question. It tracked children and teens in seven cities, interviewing them at three intervals ranging from six months to a year.

The key finding is that with participants matched for propensity, those who had contact with the police at time two (compared with those who didn’t) said at time three that they’d feel less guilt if they committed various offences from theft to violence; they expressed more agreement with various “neutralisation” scenarios (e.g. it’s OK to lie to keep yourself out of trouble); they were more committed to their deviant peers (e.g. they planned to continue hanging out with friends who’d been arrested); and finally, they said they’d engaged in more offending behaviour, from skipping classes to taking drugs or being violent. This pattern of results differed little whether police contact involved being arrested or merely being stopped. [emphasis added]

The study lends support to wishy-washy, liberal criminological ideas like labeling and “neutralization” (if you took the basic crim course, you recognized this old friend in the above paragraph). This does not mean that deterrence doesn’t work. It just means that stopping kids on a massive scale is not an effective deterrent.

The article itself in Crime and Delinquency is here, gated for $25. The summary is free at Research Digest

--------------------
*Of the 533,00 police stops in New York last year, 729 turned up firearms.  Whether a hit ratio of 0.14% is high or low is of course a judgment call.  (The police also scored 4,700 knives – lucky for me and the Swiss army that I wasn’t stopped.  Including those raises the batting average to 1 in 100.)

G.D.P. - Inclusions and Exclusions

August 1, 2013
Posted by Jay Livingston

What counts as “product” in the Gross Domestic Product?

Jared Bernstein and Dean Baker comment (here) on the new rules for calculating GDP, particularly the change that the money spent to produce “long-lived” entertainments will now be counted as investment.  These include TV shows that get syndicated (“Seinfeld” or “Law and Order”) and franchise films (“Star Wars”).  Those changes add up.  Or as Bernstein and Baker put it

the ultimate show about nothing will now add billions to G.D.P.

They also note that many entertainments that are widely produced and consumed do not get counted at all in G.D.P.  – the time people spend creating and watching YouTube videos, for example (or writing and reading blogs).

What’s really being valued here is entertainment that’s protected by copyright, which in the era of viral videos is actually a declining share of what we watch.

Later in their essay, Bernstein and Baker point out the limitations of G.D.P.

perhaps the most arbitrary part of this or any other G.D.P. revision is not the value of what’s put in, but the cost of what’s left out.

Costs like degradation to the environment.  The value of gas extracted by fracking will be added to the G.D.P. figure.  But 

there is no subtraction for the polluted groundwater or the greenhouse gas emitted when the gas is burned.

Liberals of a certain age reading this will hear echoes of Bobby Kennedy’s 1968 speech,* just three months before he was assassinated.

Gross National Product counts air pollution and cigarette advertising, and ambulances to clear our highways of carnage. . . . It does not include the beauty of our poetry or the strength of our marriages, the intelligence of our public debate or the integrity of our public officials.

The entire passage is  worth reading  or listening to.

Kennedy was speaking about GNP not GDP.  In 1968,  GNP** was the most widely used indicator.  But Kennedy’s point applies to GDP as well. They are both purely economic, with no evaluative or moral dimension.

The antidote for this non-moral measure came from conservatives – the “values” crowd.  In the early 1990s, William Bennett and the Heritage foundation created the “Index of Leading Cultural Indicators,” which did include the strength of our marriages (rates of divorce and out-of-wedlock births) as well as things like violent crime and  SAT scores.  In the next several years, with the national government dominated by Democrats, those indicators generally showed great improvement.  So did GDP. 
----------------------------
* I cannot find any information on who wrote this speech.  I suspect it was Dick Goodwin.

**For more on the differences see Wikipedia.

A Cautionary Tale

July 30, 2013
Posted by Jay Livingston

Andrew Gelman (here) summarizes and links to a “killer story” that “ is so great that all quantitative political scientists (and sociologists, and economists, and public health researchers)” should take notice.  He’s right.

First, there’s the clever finding that political moderates literally see more shades of gray than do people whose views are more extreme, right or left.  Literally.  It’s a test of color perception.*

But the more important part of the story is Part II.  The authors (Brian Nosek, Jeffrey Spies, and Matt Motyl) could have gotten the study published, but they decided to do a replication first just to put the clincher on their findings.  The result: a p = .01 effect completely disappeared: p = .59.

I’ve commented before (here) on difficulties with replication and the more general problem of diminishing effects.  (See also Jonah Lehrer’s New Yorker article  “The Truth Wears Off.”) But this is as dramatic a turnaround as I know of.

In a comment on Andrew’s blogpost, Ashok Rao suggests that authors post the odds they would give on replication.  Making the authors bet on their results “seems like a pretty good way to discern papers where authors believe what they publish from, well, that where the ‘ample incentives’ dominate.”  (Rao also links to his own paper where he quotes Alex Tabarrok: “bets are a tax on bullshit.”)

------------------------------
* If you’re thinking “50 Shades of Gray,” Andrew already beat you to it.  That, in part, is the title of his blogpost.

Anagrams - Combinatorial Probabilities

July 27, 2013
Posted by Jay Livingston

Maybe you’ve just taken a course in advanced probability.  Here’s a problem. Consider the following tweet*

What is the probability that someone else within the next day or two, coincidentally and without any knowledge of this tweet, would tweet a message that is a perfect anagram of this one? 

I have no idea even how to start thinking about it. The tweet has 29 letters, probably the more frequently used letters.  How many groupings of them form words, how many of those groupings make sense, and so on.  I give up.  But here’s one answer.

Second question.  What is the probability that someone would create a program to cull the Twitter universe, extract anagrams, and post them to a Tumblr page?  I’m not sure how to calculate that one either, but when you see the site, you might well think the probability approaches 1.0, i.e., “It had to happen.”**

This Tumblr has been up for less than a week, and so far there are about thirty examples, most of them short. It’s possible that the pool of matches has been edited to include only those that sound like they might be a conversation.  Like this:

Or this conversation between hooker_225 and FutureShrink:

You can find the entire collection at Anagramatron (here).

---------------------------------
* Ignore whatever else Victoria and Larry, with their interesting @ might be doing. Focus on the letters in the message.

** UPDATE:  My advanced probability informant tells me that it can be done with a fairly simple algorithm. Take two phrases, strip out everything but letters, sort alphabetically, and check to see if they are identical.  For the 400 million tweets in a single day, your computer has to do only about 80 trillion such comparisons.