Psychologists are often asked to testify in court cases involving eyewitness testimony. Of the numerous death-penalty cases that have been overturned in recent years, the majority of convictions were based on false eyewitness testimony. Expert testimony can be crucial in these cases. But only if the jury believes the expert. And his or her credibility ebbs and flows with the credibility of the entire field of psychology.
There is trouble brewing in psychology, at least in the public eye (here's a great summary of the issues). A year ago, a study by Daryl Bem came out showing evidence of ESP. Although subsequent studies suggest that ESP doesn't exist, the entire field's reputation seems to have taken a blow. Psychologists have recently shown that statistical shortcuts can easily create impossible results (e.g., listening to the Beatles' When I'm Sixty-four makes you younger). Psychologists are so worried about the replicability of their findings that one group plans to try to replicate every article that came out in three journals in 2008.
These efforts may undermine the credibility of psychology. And that will give comfort to a lot of crimial prosecutors. If an expert witness testifies that for example, just because a memory is vivid doesn't mean it's accurate (which is true), a prosecutor might slyly wink at the jury and say "and do you also believe in ESP, Dr. Smith?"
Psychologists understand that a quirky ESP study doesn't represent the field. And they understand that as science works itself out, some findings will not replicate, but that doesn't mean we should throw out the baby (psychology) with the bathwater.
But there is another audience for psychology research these days. In the public's view, empirical research is either completely airtight or it's just a "theory." Even evolution, which is completely airtight, is seen as debatable as long at least one "scientist" disagrees. (Here's a funny essay on the issue.) Thus, to a jury, ESP research casts doubt on psychology. Even on firmly established findings like the fact that asking leading questions can change a witness's memory. (A friend of a friend has already experienced this problem firsthand.)
Should researchers try to create a false sense that the field is unanimous about everything? Of course not. Non-replicable effects should be exposed as such, and occasional out-of-the-box research (on things like ESP) have their place. Scientists cannot censor themselves, especially to create a false impression.
But psychologists can do a better job of stating what is obvious to them, but not to the public: A new finding from last month might not be real, if it doesn't replicate, and ESP certainly doesn't seem to be real. But after decades of replication, some findings are firmly established. The justice system needs expert testimony, and juries need to believe in it.