DNA Testing and Suspect Strategies

Gregory Cochran writes:

Nowadays, inspired by programs like CSI and NCIS, many students want to become some sort of forensic scientist. The problem is that are very few such jobs. I have heard that there are something like 20 times more forensics graduates than openings. This is not really caused by an overall shortage of crimes, more by a shortage of interesting crimes. When some dirtbag stabs his old lady, after beating the shit out of her for years, caught while still gripping the bloodstained murder weapon, who needs CSI?

Indeed, one of the many big differences between fiction and nonfiction is that crimes in fiction are much more intriguing than crimes in the real world. Or perhaps I should say the portion of crimes that becomes known to the police. Students of crime like to assert that homicides are measured with little error, but nobody knows whether that's actually true, and I've been wondering for a while how many of the young children whose death is ascribed to the mysterious "sudden infant death syndrome" (a label, not an explanation) are actually homicide victims.

Be that as it may, real-world known homicides are rarely of the type with which DNA analysis will help. When a husband kills his wife, you'll hardly nail him by showing that his DNA could be found in their house. Or, closer to Cochran's example, take a case from Hanover that made the rounds in the German newspapers a while back. A German and an Italian got into a fight in a bar over whether Italy had won the World Cup three or four times (an obviously important question given that Germany has won it three times). The Italian went home, got a gun, and killed the other discussant. It seems like a bit of an overreaction on the part of the offender, especially given that he was right. Anyway, no DNA needed.

I wonder, though, whether the threat of DNA testing does help the police solve some crimes. One of the big mysteries surrounding crime is why, in countries in which you are not obliged to give information to the police once you know you're a suspect, suspects still often do. The standard police strategy seems to be to suggest to an interviewee, rightly or wrongly, that they have something connecting him (it's usually him) to the crime, and to further suggest that he can talk himself out of that connection. The possibility of DNA testing might come in handy in this respect, as it can potentially establish that a suspect was at the crime scene. And once he acknowledges that, interrogators have a foot in the door. 

Note, though, that this presupposes that presence at the crime scene is in question, so, again, this effect should be limited. I've never looked into a possible connection between DNA testing becoming available and clearance rates. At the time, though, I seemed to notice something: The typical defendant's strategy in rape cases appeared to change. Before DNA testing, it seems, defendants typically claimed that they did not have intercourse of any kind with the victim. After DNA testing, the standard strategy quickly seemed to switch to claiming that there was intercourse, but it was consensual. If my impression holds up when you look at the data, you could use these changes to estimate the something like the minimum percentage of cases which made it into court in which the defendant should have been found guilty. Someone else do it!