Time and time again, research has shown that the hiring process is biased and unfair. Factors like unconscious racism, sexism, and ageism, even the weather on the day of the interviews can influence hiring decisions.
Another study on decision-making in the United States showed that different judges presiding over identical cases meted out varied sentences. While the average sentence was seven years, identical cases had a four-year sentence disparity, the difference between a five-year and nine-year sentence.
Therein lies the flaw in human judgement. We are unable to exercise objective decision-making due to the existence of noise and the unwanted variability in professional judgments of the same problem.
Where there is noise, there is bias and more than you think
Recently, we had the honour of having Olivier Sibony, co-author of Noise: A Flaw in Human Judgement and Senior Advisor to Qualgro, speak on noise and bias at our Qualgro Symposium
Noise exists because of the existence of various factors. These include cognitive biases, group dynamics, mood, stress, fatigue, and differences in skill and taste between assessors and decision-makers.
Bias is one part of the error equation. The other is what fellow researchers in the field term ‘noise’. Noise is the “unwanted variability in human judgements of the same problem”, and it is just as problematic as bias.
Such variability leads to injustices, varied hazards, and multiple kinds of costs. Biases can shape a company or industry’s culture and norms if left unchecked. Other research has found this prevalent in various fields, from military intelligence analysis to actuarial science and virtually every industry imaginable, even one as critical as medicine.
The error equation
According to Sibony, errors can be mathematically calculated. Without getting too technical, this essentially means that the combination of both bias and noise leads to making errors, and the reduction of either has an equal impact on reducing error.
Bias and noise exist virtually everywhere, and technically, the only way to eliminate them both is to remove the use of human judgements. However, this is not tenable for myriad reasons, especially since the human element still matters in many areas concerning people (e.g., a medical diagnosis).
Noise and bias in business
As a result of noise and bias, professionals in critical industries can make important and even outrageous errors. You can see these errors of judgement in areas such as recruitment and human resources, marketing, and even when choosing which companies to invest in.
For example, a popular method of judging candidates and hiring them is based not on objective data but on the gut instinct of the interviewer. Unfortunately, such decisions can lead to bad outcomes and incur extra costs for the business.
The next best thing would be for humans to learn how to reduce errors in their decision-making, especially with cognitive biases.
Decision hygiene factors
How exactly do cognitive biases and noise affect entrepreneurs, and more importantly, how can we reduce decision-making errors within the business landscape?
Sibony references decision hygiene factors, a matrix comprising four noise prevention techniques to help make better judgements and decisions.
In some specific situations, a diversity of input can be useful in the decision-making process, so long as the inputs are independently derived. aggregating independent inputs and then averaging them out would statistically reduce noise.
- Use relative measures, not absolute
Chances are, when people describe things or situations, they will use the same terminologies despite meaning different things. This can be problematic if two people use the word “great” to describe what sort of potential investment needs to have, but Person A means 30 per cent while Person B means it is in the seven per cent.
Since absolute measures can be ambiguous, it would be better to rank or measure items or situations against others. For example, before deciding to invest in a start-up that, say, sounds great on paper, compare them to other start-ups similar in scale and size for their relative performances.
- Structure your judgements
Break judgements down into separate components or dimensions and use quantitative and objective measures to assess and/or analyse sub-components of the judgement you will be making, and score them against a frame of reference.
For example, you can structure an interview process to have several stages where specific competencies are assessed (such as through scoring) and compared to other candidates similar to them (relative measuring).
You can then aggregate independent inputs and average their score or performance.
- Keep intuition at bay
Humans are generally susceptible to cognitive biases such as selective attention, confirmation bias, and selective recall.
This can make you over-focus on some types of information and overlook other relevant ones, leading to terrifying outcomes. Just ask Brandon Mayfield, who was wrongly detained for the Madrid bombings.
“The key point here is that you don’t want to know what you don’t want to know,” quips Sibony, “knowing too much, even accurate information, can mislead you”.
Manage the information process to make it difficult to form an intuition too early. Although it is tempting to engage in intuitive judgements, early use of intuition only serves to add more noise.
Indeed, many people think they are very objective and impartial, especially when their professional judgements are solicited. However, as illustrated, erroneous judgements and decisions can lead to disastrous consequences.
As executives, the company and the organisation depend on not just the knowledge and experience but, more importantly, the sound judgement and decision-making skills of executives.
It pays to be cognisant of how we may stumble at different stages and work towards strengthening noise-prevention efforts for the health and success of the organisation.
This article was originally published on e27: https://e27.co/avoiding-costly-mistakes-how-cognitive-biases-can-affect-entrepreneurs-20220914/