Anchoring Your Biases

I couldn’t tell you when I last listened to a football commentary on radio or TV (the current World Cup has not caused me to rush to change this), but I don’t need to listen to realise that a statement that ‘women’s voices are too high’ to do a decent job at the commentary is liable to be perceived as inflammatory. Why, I ask myself, should the pitch affect the accuracy, the shrewdness or the insight being proffered?  Maggie Thatcher famously dropped her voice to sound more authoritative (it would seem successfully) and I too was once advised to take voice coaching lessons so I sounded more ‘persuasive’. I felt then, and still do, that I would rather talk sense in a high voice than drivel in a low one, and would like to listen to others talking sense too whatever the pitch of their voice (or accent come to that).

But that isn’t the point I want to make here, it is the way that statement assumes there is a default benchmark. This tactic is all around us.  I recently saw an ‘advert’ on a bus, I assume for CRUK, saying something along the lines of ‘we will beat cancer sooner’. Clearly a statement that at first sight it seems hard to disagree with – but sooner than what? Really what they mean is that with CRUK funding it may be possible to do something about some cancers (which ones?) a bit earlier than if that funding was not present.  A laudable aim but slightly less dramatic than their tagline. Perhaps because of cancer having come too close to my own family, their claim (or another one, embossed on the wall of a CRUK building saying ‘together we will beat cancer’) I just find antagonises me.

Or how about all the imprecise statements we read on the web along the lines of ‘adding an extra portion of unprocessed red meat to someone’s daily diet would increase the risk of death by 13%’. Meaningless statistics – we’re all doomed and going to die anyhow – but it is the relative statements that can sound so powerful and yet mean so little. Our media, and apparently our pundits, are full of them. The trouble is, implicitly they all assume some baseline to which everything else is referred. Sometimes this is obvious, but too often it is implicit rather than explicit. That a man’s voice carries more gravitas than a woman’s is, apparently, to be taken as read. That this is so has been presumed since Classical Times, and the words used to describe women’s voices – shrill, strident, bossy – convey that negativity. That a person is a man and not a woman – who traditionally needed to be labelled by her marital status whereas a man could just be referred to as Bloggs (unless he was a member of an order of Chivalry or the Aristocracy when presumably Lord Bloggs would have been more appropriate) – was also taken as default. For some of the UK’s history a woman was, legally, quite literally a non-person. It isn’t that long ago, certainly within the last decade, that in Cambridge committee membership lists had the identifier (f) after a woman’s name, whereas there was at that time no such qualifier of (m) against the men. Male by default – until I made a fuss.

Anchoring to a baseline, be it life expectancy or gender, happens pervasively in our society and its dangers lurk in its very invisibility. Tversky and Kahnemann identified it in the context of choice, when how you frame a question affects the way people respond. Initially the construct was economic: would you be willing to take a punt on a choice when the question was framed as a possible loss from some sum, as opposed to a possible gain from some lower sum? Even if the end position was identical, people were more likely to go for a possible gain than risk a possible loss. However, it applies equally to selection processes: is this person you might appoint more or less like the one you have in mind/in post? And, carried to an extreme that actually isn’t that far-fetched, does their voice convey the gravitas you feel the role needs? In that sense its pernicious effect is prevalent in our universities where unconscious bias of this sort frequently raises its ugly head.

The trouble is that unconscious bias ‘training’, certainly if taken as a tickbox exercise achieved simply by watching a 3 minute generic video, does not necessarily work. Indeed, some evidence suggests it is worse than useless by making people feel virtuous and that they ‘get it’ to the extent that they don’t need to consider it further. For some people it may cause reflection, for others it may just make them yawn. A report published earlier this year by the Equality and Human Rights Commission explores the efficacy of training methods. It should give us all pause.

In my university, the training that I, as a member/chair of an appointment’s committee, am required to do consisted of an afternoon group session where we were all expected to contribute our prior experiences – good and bad – and to reflect on what the pitfalls were.  This is a very different and, I would like to believe, much more effective way of informing each and every one of us of the wide variety of ways in which unconscious bias can raise its ugly head, of which anchoring bias is just one. It is a start to be able to share examples that won’t have crossed everyone’s path (and these extend to many characteristics way beyond gender). I would say it is the time devoted to reflection and discussion that impacts on changing behaviour, not being told some bare facts that you don’t necessarily absorb, or wish to do so. In that sense I have moved on from a stance I suspect I held a few years back, that give everyone the training (i.e. that brief video) and the problem will go away.

There is no doubt that I have seen far more instances of interventions around bad behaviour in recent years: the man who points out that a double standard is being applied to a field of men and women, and everyone agreeing and being more careful; the Chair who sets out the rules before a meeting ever gets underway to ensure that everyone is up to speed, which might be about how impact factors are to be ignored in every case and not just to suit someone’s favoured candidate; the process of drawing up separate shortlists of men and women which are then interwoven, to help avoid unconscious bias at the earliest stages when it would seem women too easily get dropped without proper consideration. All these strategies may help to ensure the playing field starts to approximate to level. There is a way to go to ensure everyone gets as fair a chance as possible, without some extraneous benchmark being applied just because it feels comfortable but has no bearing on the actual situation or job to be filled. But the more these issues are discussed, the greater the likelihood that we will reach a better place.


This entry was posted in Equality and tagged , , . Bookmark the permalink.

One Response to Anchoring Your Biases

  1. Laurence Cox says:

    Yes, that BBC web page is a shocker. Anyone who has read Ben Goldacre’s “Bad Science” would know that the best way to describe statistics like these is by the use of natural frequencies (e.g. 1 extra death per per year). If it is an increase from one in a million to two in a million, that is a horrifying 100% increase in relative risk, but still overall only a small increase in the death rate.

Comments are closed.