We all know that bias is a problem for scientists and we know that in order to design a good study, a scientist must find ways of eliminating potential biases where possible. But bias is a problem in many areas – even if it sometimes seems that the only ones who care are scientists. Selection bias is something that scientists must be careful of when selecting subjects or looking at what studies or data to use and recall bias is a common problem when using human subjects (say, when asking people about their diet or the amount of exercise they take in order to find out more about relationships between nutrition and health or between exercise and health). Recall bias can often be a problem in certain other fields.
One possible example of recall bias is from the Vanzetti trial of 1920. The Pinkerton detective agency interviewed four witnesses to the botched hold-up Vanzetti was later found guilty of, three of whom had been in the payroll truck when the attempted robbery took place. The following are summaries of the descriptions given by the witnesses to the Pinkerton agency. Paymaster Alfred Cox described the suspect thus: 5’8″, 150 pounds, dark complexion, 40 years old, closely cropped moustache which might have been slightly grey. Didn’t identify the car. Driver Earl Graves gave this information: dark complexion, black moustache and looked like a Greek. Believed the car was a Hudson. Police officer Benjamin Bowles and bystander Frank Harding were the other witnesses interviewed. Bowles: black closely cropped moustache, red cheeks, slim face, black hair and was either Italian or Portuguese. Offered no description of the car. Harding: 5’10” and thin. Did not get a good look at his face but thought he was a Pole. He identified the car as a black Hudson #6 auto.
So far then, we have a suspect who is 5’8″ to 5’10” with dark complexion and a black, closely cropped moustache, of Greek, Italian, Polish or Portuguese nationality. We also have a car that is not identified at all by two of the witnesses and is described by the other two as a Hudson. An Italian man named Vanzetti with a big, bushy moustache (and who had been linked to a Buick owned by another Italian man) was later indicted with this crime. Here’s what happened next.
At the preliminary hearings, Harding (who “did not get a good look at the suspect’s face”, remember – and neglected to furnish any description of the suspect’s face to the Pinkerton agency) decided that the man had a dark moustache and high cheekbones and was ‘swarthy’, adding that there was “no question in [his] mind” that Vanzetti was the one (he “didn’t get a good look at his face” originally – how the hell can he be in no doubt as to the man’s guilt?). He also changed his description of the car from a “black Hudson #6 auto” to… a Buick. Bowles, meanwhile, reaffirmed that the suspect had a short, croppy moustache – but then pointed to the heavily moustachioed Vanzetti (whose barber later testified in court that he never cropped Vanzetti’s moustache – it was always big and bushy). Cox was asked to identify the man and asked the court “how do you want it worded” he went on to say “I think it is this man behind the rail, the man with the moustache” and finally “I think he looks enough like the man to be the man”. It’s meant to be an identification of an actual person, not a Tom Jones look-a-like contest at Butlins.
At the actual trial Cox gave more details when describing the suspect, including the ‘high cheekbones’ that had appeared in none of the original descriptions – but did appear in Harding’s revised description at the preliminary hearings. If Cox failed to give this detail on the first two occasions on which he was questioned, how was he suddenly able to add it to his description now? Bowles, who had up until this point remained silent as to the type of car involved (and remember, this is the third time of asking) suddenly remembered it as a dark-coloured Buick. Which is in line with Harding’s revised description of the car as given at the preliminary hearings – but does not tally with the two original descriptions of the car as being a Hudson. Harding had originally been fairly specific in regards the car used in the failed heist and said it was a black #6 Hudson. This black #6 Hudson had become a “large car, dark car either blue or black, I should say, a seven passenger Buick” by the time of the trial. Bowles originally knew nothing about the car, Harding (like Graves) thought it was a Hudson. Both their stories changed and the car became, in both cases, a Buick. Bit of a coincidence, that. The jury was never made aware of the preliminary hearing or Pinkerton witness statements and the Pinkerton statements were only made known to the defence in 1927.
A suspect’s moustache being ‘bushy’ rather than ‘croppy’ and a misidentified car might not seem all that significant, but the identification of Vanzetti by these witnesses and the identification of the car by these witnesses were important for the prosecution case. Vanzetti was convicted of the crime and sentenced to 15 years imprisonment. Following a related criminal trial, he and a man named Sacco were sentenced to the electric chair – by the same Judge who had presided in the 1920 Vanzetti trial. They were executed two months after the Pinkerton evidence was made known in 1927.
Judge Thayber, who (according to Lee and Labriola) was noted for xenophobia, had actually written to the Chief Justice and requested that he be given the appointment. He also presided over all six unsuccessful appeals. Vanzetti was under indictment for the second case when he was tried and convicted in the first case. The fact he was under indictment for the second case (a robbery-murder) may have been a factor in the outcome of the first case. The fact he was already convicted of an attempted robbery may have been a factor in the second case. Given that the same judge presided over both trials, we can be sure that whether or not he was affected by these factors he was most certainly aware of them. Vanzetti and Sacco were apparently anarchists. This being the 1920s, the Americans were just a tad anxious about communists, socialists and anarchists. I think this provides some useful information as to the context in which this case was seen by Americans at the time. Particularly those chauvinistic, patriotic Americans like lawyer Barry C Reed who described the defendants as “aliens, anarchists and draft dodgers”.
I think this case is a good example of how people and organisations can (whether by accident or design) fail to eliminate bias and thereby ensure that there will always be doubt as to the reliability of their findings. There is still doubt as to the guilt of Sacco and Vanzetti and there is still interest in the trials. I know this because I was reading about it in Famous Crimes Revisited by Drs Henry Lee and Jerry Labriola earlier today. Frankly, I preferred the Michael Baden book about autopsies, but I did find the inconsistencies in the Sacco-Vanzetti witness statements fascinating. Also of interest is that the witness statements of 14 Italians who saw Vanzetti delivering eels on the morning of the botched robbery seem not to have been given the same credence as the contradictory and ever-changing statements of four Americans.
Links: The book is on Amazon and there is a page on Sacco-Vanzetti on Wikipedia. EDIT: There’s more here – Sacco e Vanzetti [includes letters written by Vanzetti and comments made by him to journalists] and there’s a Washington Post review of a book about the case here. Also, DrAust has left a comment that included this link. Yet more here.