Time In and Time Out

Maybe it's time to get soft on crime.

That's because many criminals are more likely to go astray once they get out of prison if they faced longer sentences and more punitive conditions in the slammer, claim economists M. Keith Chen of Yale University and Jesse M. Shapiro of the University of Chicago.

"Harsher prison conditions are associated with significantly more post-release crime," they report in their updated working paper posted on the university Web sites, a finding that suggests doing hard time often may only produce more hard-core crooks.

For their study, Shapiro and Chen looked at convicts with virtually identical criminal histories and examined the "security risk" score each federal prisoner is given before entering prison. The rating, which ranges from zero to 36, is based on the prisoner's rap sheet, predisposition to violence and other factors. (The score determines whether an inmate is assigned to a "minimum-," "low-," "medium-" or "maximum-" security prison.)

The researchers focused on inmates who had ratings within a few points of each other but were assigned to different security levels because they were just under or over a cutoff. Chen and Shapiro reasoned that roughly similar criminals should have roughly equal probabilities of committing crimes once they were released.

Scratch that theory: Offenders who scored barely under the cutoff point and served time in a minimum-security environment were only half as likely to commit crimes in the three years after release as those unfortunates who scored just high enough to be sentenced to the next-higher security class. The same general pattern appeared to hold true at other cutoffs.

So why were those crooks who did harder time twice as likely to get into trouble again? Shapiro and Chen suspect that those who fall into the higher security class are housed with more hard-core, violent criminals who may school them in the dark arts or otherwise encourage them to resume their lives of crime.

The Experiment: Let's Break Some Ground

---

What's important to people when they think about political candidates? Here's your chance to help us find out and have a little fun while participating in groundbreaking research in the social sciences. Unconventional Wisdom, washingtonpost.com and Stanford University's Shanto Iyengar are collaborating on a series of online experiments similar to those he pioneered at the Political Communication Lab at Stanford. These are serious studies, designed to test real hypotheses on important issues in political science, sociology and psychology.

We're inviting you to participate in the inaugural experiment. If you're interested and promise to take our study seriously, go to the link below and read a description of the experiment and its purposes. Then you're ready to start; it won't take long. In a few weeks, we'll report the results in this column and on the web.

So let's begin. Go to http://www.washingtonpost.com/richmorin , look for the box headlined "Post-Stanford Experiments" and click on the link.

9/11 Traffic Effect

---

A grim but unexpected consequence of Sept. 11, 2001: Traffic fatalities spiked after the terrorist attacks on the World Trade Center and the Pentagon as thousands Americans temporarily abandoned air travel to take their chances driving.

Daniel H. Simon and two colleagues from Cornell University estimate that highway deaths increased by about 242 fatalities a month in the last three months of 2001 in response to the terrorist attacks after controlling for time trends, weather, road conditions and other factors associated with highway accidents.

In all, they estimated that "at least 1,200 additional driving deaths are attributable to the effect of 9/11" through 2002, when fatalities gradually returned to usual levels.

"Our results show that the public response to terrorist threats can create unintended consequences that rival the attacks themselves in severity," economist Simon and his colleagues report in their paper available on the Social Science Research Network.

Who Would Have Thought? Love Hurts, Dogs and Boy Toys

---

* "An Examination of Sexual Strategies used by Urban Southern and Rural Midwestern University Women," by Peter B. Anderson. Journal of Sex Research Vol. 42 No. 4. A University of New Orleans researcher finds that women attending a Southern university appear more likely to have hit, held down or otherwise used physical force to get a reluctant boyfriend to have sex with them than coeds at a rural Midwestern college.

* "Dogs Still Do Resemble Their Owners," by Michael M. Roy and Nicholas J.S. Christenfeld. Psychological Science, Volume 16, Number 9. Two University of California at San Diego researchers defend their landmark study that found that people shown photos of purebred dogs and then photos of dog owners are able to match the owners to their pooches with uncanny accuracy.

* "Action Figures and Men," by Chris Barlett et al. Sex Roles: A Journal of Research, Vol. 53 No. 11/12. A Kansas State University researcher and his colleagues find that male test subjects who played with unrealistically muscular toy action figures such as "Battle Action Hulk" for 30 minutes suffered diminished body self-esteem.