Dr David Schudel forensic chemist at Keith Borer Consultants.
According to the Oxford English dictionary, “bias” has several meanings, including: “to influence, affect (often unduly and unfairly)”.
No one likes to think of themselves behaving unfairly, and no reputable expert would ever regard themselves as anything other than a neutral advisor in an otherwise partisan system. This is why conversation about “cognitive bias” can be difficult: despite cognitive bias being innate, unconscious and normal, the topic is still uncomfortable because no one wants to admit to being “biased” in any form, even if they can’t help it. So let’s try calling it “subconscious influences” and see if that makes the subject more palatable.
Subconscious influences were recognised a long time before the term “cognitive bias” was introduced to our parlance. Four hundred years ago, Francis Bacon in his Novum Organum (1620) talks about how the human mind, once it is fixed on a thought, will ignore, despise or reject anything to the contrary “with violent and injurious prejudice, rather than sacrifice the authority of its first conclusions.” Despite this, and the heavy reliance on the human mind in Forensic Science, it is only recently that the concept of cognitive bias entered the forensic arena.
In 2015, the Forensic Science Regulator (FSR) produced, “Cognitive Bias Effects Relevant to Forensic Science Examinations, FSR-G-217”; a document that received its second and only revision in 2020. FSR-G-217 illustrates the different forms of subconscious influence that exist and how they might come into play during a criminal, or indeed any, investigation. Despite this document approaching its eighth anniversary, examples of some of the more common cognitive influences continue to be seen in casework.
Take, for example, a recent case where a defendant was charged with possession with intent to supply on the basis of results from an accredited drugs laboratory. The offending substance was submitted to the testing laboratory on the assumption that a drug was present and when results came back showing the presence of heroin, ‘expectation bias’ took hold, i.e. the prosecution team took at face value that they had what was expected even though heroin in the UK is typically brown, not white.
The defendant steadfastly denied the allegations, stating the item was glucose… and in fact it was. Despite being analysed using an accredited method, two human errors at the laboratory produced a result that was wrong.
Expectation bias often goes hand-in-hand with another subconscious influence: confirmation bias. This is where a person will ignore anything that contradicts their theory and look at those facts or details that support it, rather than, as Bacon said, “sacrifice the authority” of the first conclusion. In the drug case mentioned above, the alleged heroin was a fine white powder and that alone contradicts the notion it was heroin. Despite this concern being raised by the defence, the drug laboratory was never asked to review their conclusion; i.e. the prosecution team were so convinced that the tested substance was heroin that they chose to ignore the fact that it didn’t look like heroin. Instead, it required a review of the laboratory’s data file by me to spot where the chemical analysis system had been programmed incorrectly causing the wrong sample to be analysed.
Anchoring bias effects are another subconscious influence that the FSR talks about. This is when a critical event at the start of a case anchors the investigation down a certain path and then what follows are confirmation and expectation effects that keep the case anchored to the same initial assumption. In crime, this can start with someone seen on CCTV (and flagged as a suspect); a fingerprint or DNA database hit; or a convincing eye-witness account.
By way of example, in one of our previous cases, fingerprints that matched a person on the fingerprint database were found at a school break-in. This person lived in the area, had a criminal record and was consequently charged with the crime. The accused person denied having been to the school and offered that he worked for a window manufacturer 15 years earlier. The school staff said the windows were regularly cleaned and on the basis of this, the ‘previous contact’ defence was discounted – the anchoring bias being that the fingerprint must belong to the perpetrator of the crime.
However, “cleaning” is a subjective term and there’s actually no research that can tell you how long a fingerprint will last on an indoor window frame. In this case, our analysis of fingerprint orientation (as opposed to source, which was all that was considered by the police) showed it was impossible for the marks to have been made after the window was fitted. The only plausible explanation for the findings therefore was that the fingerprints were 15 years old!
What are the chances of that? I’m not sure, but we’ve had 3 other fingerprint cases like it. As our fingerprint specialist, Simon Bunter, often asks, “What did you touch 15 years ago? If something ends up on a database and gets a match years later, will you remember where you’ve been?”.
There are other common types of subconscious influences discussed in FSR-G-217, and whilst the document gives a good overview, there are, in my view, two notable omissions:
- “Status quo bias,” which is really “you can’t teach an old dog new tricks”. For example, reporting of fingerprint comparison work has remained largely unchanged for 120 years despite several studies, mainly out of the USA, showing reoccurrence of false positives in controlled tests – something that in the very least should require an error rate to accompany a ‘match’ statement.
- “Accreditation bias” where, for example, it is believed an expert using accredited methods must be correct. This will no doubt have been a factor in the drug case referenced earlier and is becoming an increasingly common. In our view, this is sometimes a misplaced justification for one expert’s opinion being correct and another’s not. Bear in mind that it is a common misconception that people get accredited, but they don’t; it’s the method that is accredited. It should go without saying that using an accredited method does not guarantee quality of opinion.
In many ways, it is unimportant what type of subconscious influences are in play – what is important is to understand they happen – it’s normal, they affect everyone. As ‘defence’ experts we are in the privileged, and often unique, position of being able to access the full range of scientific findings and assess them in the context of all of the case information. That is not to say that we are immune from subconscious influences; after all, we’re only human, but at Keith Borer Consultants we are acutely aware that they exist and do our utmost to protect against them. Critical findings checking and effective peer-review are important parts of this protection, and we would urge caution if you are considering engaging the services of an expert who employs neither of these measures.
If you have questions about forensic work done in your case, please get in touch.
Dr David Schudel is a forensic chemist at Keith Borer Consultants.