“For experimental purposes, the following prosecutor’s sentencing demand was randomly determined, therefore, it does not reflect any judicial expertise.”
That sentence was printed in a questionnaire handed to 39 German legal professionals at a training conference. Thirty-seven of them were judges, and on average they had spent more than 13 years in court. They had just read a case file about a woman, called Lena M. in the materials, caught stealing from a supermarket for the 12th time. Their task was to sentence her.
Under the warning came the number. Half of them read that the prosecutor wanted 3 months on probation. The other half read 9. The defense always asked for 1.
The psychologists Birte Englich, Thomas Mussweiler and Fritz Strack had already asked 13 other experienced lawyers to sentence the same woman with no demand in front of them. They averaged 5.62 months. The paper’s first line states the principle they all worked under: sentences should be guided by facts, not by chance.
The judges who saw 9 gave her 6.05 months on average. The ones who saw 3 gave 4.00. Both groups said they felt fairly sure of their decision, and how sure they felt had nothing to do with which number they had seen.
The researchers still had one doubt. Maybe the judges had not really believed the number was random. So in the next study they put the randomness in the participants’ own hands.
Fifty-two young lawyers at a postgraduate program in Speyer, recently out of law school and in their first jobs as judges, read the same file. They were told the study was about the best order of questions in a courtroom. Then each one picked up a pair of dice, rolled, added the two faces, and wrote the total in as the prosecutor’s demand.
The dice were loaded. One set always landed on 1 and 2. The other always landed on 3 and 6. Those who rolled 9 gave 7.81 months. Those who rolled 3 gave 5.28. For the identical theft, sentences in the room ran from 1 month to 12.
Experience did not help. In the two studies with seasoned professionals (the other one used a rape case and a journalist’s phone call as the stray number), specialists in criminal law were pulled by the number about as much as lawyers from other fields. The difference was confidence: the specialists rated their certainty 6.88 out of 9, the others 4.45. The authors wrote that expertise did make a difference, but one “only apparent in the subjective not the objective quality of the judgment.”
A fourth study looked at why. After a high roll, participants sorted incriminating arguments about the case faster than after a low one.
These were small groups, a few dozen people each. In 2021 a meta-analysis of 29 legal anchoring studies, with 8,549 participants in all, found the effect again, with some signs that published studies overstate it, and found professionals anchored a little less than laypeople, by a margin too small to be statistically significant.
Judges do not usually roll dice before sentencing, as the authors noted. They pointed instead to numbers from the press, from visitors to a hearing, from a judge’s own family. Their last line borrowed from Einstein and turned it around: judges “may unintentionally play dice with criminal sentences.”