with No Comments

Post No.: 0277substitution


Furrywisepuppy says:


In such a complex world, it’s remarkable that we rarely ever feel stumped, for we have intuitive feelings and opinions about almost everything that comes our way! We like/dislike people long before knowing much about them, we trust/distrust strangers without knowing why, we feel a venture is ‘bound to succeed’ without really analysing it. Whether we overtly state them or not, we often have answers to questions that we don’t completely understand, and rely on evidence we can neither explain nor defend.


One main reason for this is that, when faced with a difficult question – if we don’t have enough information or a satisfactory answer is not quickly forthcoming (and/or we cannot be bothered to use the information that’s available or do more research via our reflective system two, or simply admit that we don’t know the answer) – our intuitive system one typically answers an easier proxy or surrogate question that’s (perceived to be) related instead; usually without noticing this mental ‘attribute substitution’ or simply ‘substitution’.


For example, if one doesn’t know whether a job applicant will be competent or not, one may answer the question ‘is he/she likeable or charming?’, ‘does he/she interview well?’ or ‘does he/she look competent (according to the shape of his/her face, posture, age or dress sense)?’ instead. And one will give an answer to the intended difficult original question based on the answer of the easier substitute question.


A classic substitution is judging the outside of someone/something when we really want to judge the inside of them/it e.g. judging a person’s character by his/her appearances or a product according to its packaging. If we want to judge how strong someone is, we might judge them by their size and/or gender. People sometimes substitute the intended question of ‘how intelligent is he/she?’ with the superficial question of ‘how much does he/she earn?’ Substitution typically involves using superficial judgements in place of intended deeper judgements. We cannot reliably judge a lot of important attributes via appearances… yet we’ll still try. Woof!


Most of the time this process is unconscious and unintentional, but we can sometimes notice a person actively disregard the original question asked of them when they start to find the answer to an easier substitute question they’ve just asked themselves e.g. they might say, “Well he/she looks trustworthy.” We judge something else but believe we’ve judged the target attribute, and it’s not due to a misunderstanding of the question nor a belief that the questions are synonymous.


Now the answer to the substitute question isn’t necessarily simpler or more frugal per se – the answer to it is only more accessible and computed more quickly and easily. Our mental shortcuts or judgement heuristics include similarity and representativeness, attributions of causality, and evaluations of the availability of associations and exemplars/typical examples/prototypes. We tend to overly judge using our eyes and judge based on what emotions we feel about a subject, as a substitution to what we’re trying to really judge – the dominance of trying to reach an answer or conclusion quickly, over making carefully considered arguments, is most pronounced whenever emotions are involved (the ‘affect heuristic’).


Evaluating people as attractive or not is automatic whether we want to or not, and it greatly influences us. We use how we feel right now to judge how we feel generally. We use past performance to judge how we think something will perform in the future, when all this can be very unreliable. The affect heuristic can make us believe that a project’s costs are low and its benefits are high when we like the project, just like it can make us believe that a person’s fuzzy foibles are minor compared to his/her positives when we like a potential date. Conversely, if we dislike a notion, such as governmental meddling, then we’ll likely reject, with little deliberation or reasoning, government interventions that, for instance, even improve our health or the environment.


So this system one, or lazy system two, mental shortcut is a simple procedure that helps us to find quick, low-effort, adequate or at least acceptable, although often imperfect, answers to difficult questions. (We can employ effortful system two shortcuts, such as to solve certain maths problems e.g. a trick for multiplying any number by 11, but they are different to these automatic system one or lazy system two shortcuts.) We intuitively favour quick answers over effortful considerations. Trying to deal with complex questions seriously is sometimes impractical though and requires such a problem to be broken down into smaller questions – but system one will only consider one or two of those smaller questions rather than systematically try to uncover and weigh them all out; and system one will usually focus on the affective and/or superficial proxies too.


These heuristic alternatives to careful and thorough reasoning frequently work well enough, but at other times they can lead to serious errors of judgement. We should slow down to let system two intervene to check or reject an answer that system one has offered. However, since system two is lazy – whatever system one proposes tends to be endorsed by system two without scrutiny or modification. This is why we’ll seldom feel unopinionated despite living in a highly complex world. You’ll not have to mentally work very hard, you may not notice you didn’t really answer the question you were asked, and you may not have even realised that the real question was difficult and complex – because a biased and intuitive answer came readily to mind. And we seldom question our intuitions.


A lot of people in fact think it’s smart of their intuitions to give answers to questions such as ‘what kind of father do you think he’ll be from the way he looks?’ or ‘how good a politician do you think she’ll be from her fashion sense?’ But the actual smart or best answers to these kinds of queries are to ask more questions and seek for more relevant information before reasoning towards any opinion, or to reserve one’s judgement if not. The reasonable and wise answer is often, “I do not know” based on the limited relevant information supplied and/or the difficulty of trying to predict the future in certain fields. A lot of people unfortunately trust those who seem the most certain of their answers though, when we know from countless experiments that confidence does not correlate with correctness!


The order of questions can affect the answers given e.g. if one had just answered positively to the question ‘how many dates did you have last month?’ then one will likely also give a positive answer to the question ‘how happy are you these days?’ because of the emotional priming and substitution of using the answer to the easier first question to help answer the more complex but associated second question. But if the order of the exact same questions were reversed, the correlation between the answers will likely be low because dating is not the main thing we tend to think about when trying to answer such a global question of general happiness.


When only asked a standalone question of ‘how happy are you these days?’, we tend to answer the substitute question ‘how happy am I right now?’ This priming effect is however not always reliable e.g. an explicit priming of the weather can provide us with an explanation for our current mood, thus undermining the connection that would normally be made between our current mood and generalising that to our overall happiness.


Consciously ask yourself as an interviewer – do small spelling errors on a CV/résumé really matter that much to the job position that’s actually on offer? Spelling and the visual appeal of a CV/résumé frequently become substitutes to what we really want to answer, which is ‘how well will this candidate do?’, when we don’t have enough direct and relevant information about how well they’ll likely do in the role in question. Judging people based on their social media pages is also an indirect, proxy judgement, often of things that are highly contextual e.g. yes, the applicant got pretty crapulent on his/her 18th birthday but will this likely mean he/she’ll get drunk at work? Once again, it’d be a case of answering a substitute question (how they are when they’re having fun on special occasions) rather than the question we really want answered (how they’ll be in this job from Mondays to Fridays).


It’s as unreliable as judging a person by how they look when we don’t yet know much else about them. (Substitution was mentioned in Post No.: 0159 regarding our first impression judgements.) An astute job applicant will nonetheless care highly about any spelling errors on any submitted materials and what sort of image he/she may be portraying on his/her social media pages; but at the same time, an astute employer will be able to put these into context, and look for ways that more relevantly reveal and assess the candidate’s potential competence in the actual role they’re expected to perform e.g. via directly relevant tests or trial periods on the job. So perusing a person’s social media pages can reveal some useful information about them but we’ve got to ask ourselves if something is relevant or are we committing an attribute substitution?


A lot of qualities can be judged along a scale of intensity, and system one allows and uses ‘intensity matching’ across diverse dimensions e.g. if crimes were colours, murder would be a deeper shade of red than theft, or if punishments were sounds, the death penalty would be far louder than a fine. We’d also expect the intensity of the punishment to match the intensity of the crime. We find it easy to answer ‘how furry is a chinchilla who is as furry as Alan Turing was clever?’ or ‘how much money would I put in according to how much anger I feel about sexism?’ But this mode of prediction by intensity matching, of translating one scale to another, is statistically wrong because the dimensions are completely unrelated/independent or are only weakly related/dependent, even though it feels natural to system one to be able to do it. If you think we can objectively convert, say, ‘one unit of joy’ into ‘one unit of temperature’ then it’ll erroneously feel acceptable to system two too.


System one also sometimes computes far more than what we want or need, akin to an imprecise ‘mental shotgun’ approach e.g. answering whether one thought a firm was financially sound but not being able to forget that one liked their products or hated their CEO i.e. we find it hard to ignore or discount irrelevant information when only asked to answer a specific question or evaluate a particular attribute.


So even some intentional assessments aren’t completely under our control – it seems difficult for our unconscious to not do more than what our conscious self instructs it to do. The intention to perform one computation can evoke another e.g. if intending to see if two words rhyme, such as ‘night’ and ‘kite’, we might also compute if their endings are spelled alike, even though the spelling is irrelevant here. The combination of a mental shotgun approach with intensity matching also contributes to help explain why we have intuitive judgements about many things we actually know little about. They both allow the substitution heuristic to work i.e. answering something that wasn’t really asked.


In summary, the attribute substitution bias is when trying to make a judgement about something that’s hard or effortful to calculate, we substitute it with a more easily accessible attribute instead, such as judging intelligence via someone’s age, skin colour or another differentiating factor, or whatever a person’s height or being bald-headed are supposed to imply about them. But wearing geeky glasses doesn’t make one more intelligent, drinking protein shakes won’t alone make one more healthy or strong, wearing pink doesn’t make one gay, and so on…




Comment on this post by replying to this tweet:


Share this post