Our Systems are full of noise, which means a judgment based on assumptions, not on facts. Noise and bias can engender real problems for example in the Judiciary system, many studies prove that judges were exposed to the same cases but announced different sentences which varied a lot from each other.
Many factors can influence judges and consequently their sentencing such as mood, temperature, and other judges' opinions: Judges are more likely to be lenient on their birthdays, when their favorite team won, or when the weather is good. It seems that sentencing should be more serious than that so you are better off being lucky enough to get a happy judge. The difference between two judges judging the same cases can be many years so it's obvious that noise can cause unfairness.
On the other hand, studies demonstrate how important are guidelines and hiring a noise audit to reducing noise because it seems that people aren't aware of how much noise exists, when they were asked, they underestimated the quantity of it because they were so confident about their judgments to the point they think that others adopted the same thinking, that appears right for everyone for the simple reason: they got the same training and instructions.
Crowd-wise is another method to reduce bias and noise among people, many references show that the average opinion of a wise crowd(experienced community) increases the chances of making the right judgment or even the right prediction. Another method is to choose good judges, judges who are experienced, think well, and differently. People who get a characteristic called respect-expert, they have good reputations, they are respected by their peers and they're well trained. There are many other cognitive traits that a good judge should possess such as decision-making, intelligence, a high scale of GMA, an open mind to learn anything new, and sometimes accepting counterarguments. I understand that leaders have to look confident, as they know everything in their deep bones but it doesn't mean not crossing the data given to you and looking for new information.
Algorithms and following common guidelines can contribute positively to reducing noise in Judiciary systems but we still don't trust technology, we love to trust our gut or what we call intuition, we do and love to give everything a reason why happens:we base our thinking on memories, the way we are taught how to explain events, to tell stories about how should things get done or be but what if the outcomes are good, in this case, we don't look after the reasons because they're obvious to everyone.
Debiasing can also be helpful but we should distinguish between two types:
Post-debiasing: it necessarily means correcting the biases that already exist or even influencing the outcomes and freeing them from bias.
Ex ante: On the opposite hand, it essentially prevents bias in judgments by taking diverse measurements and using many methods (the methods we mentioned above).
In ongoing paragraphs, I will treat other methods in two different fields: the HR department and the forecasting department.
Let's start with forecasting, forecasters are more exposed to noise and bias because their mission is to predict phenomena that not happen yet famously the weather and natural catastrophes.
Many studies prove that forecasters are aware of the existing noise and they're better at reducing it, unlike the judges who ignore it most of the time.
It exists three ways to improve judgments in the forecasting field:
Training: more training equals more knowledge so well prepared and ready to face noise in judgments.
Selection: to select good forecasters, the smartest and the most experienced ones.
Teaming: when experts aggregate their thoughts and opinions, the chances are getting bigger in reducing both noise and bias by crowd noise effect.
The HR department's function is to hire people who fit into the company and can contribute to its future and goals.
The HR department in most companies use often unstructured interviews, which means asking multiple questions and getting very variable impressions but generally, the first one impacts the others (Cascade noise). After all these studies showing that noise exists more than we think, Google decided to process interviews differently: in the strategy of structured interviews, the hiring process went through three steps :
Decomposition: in this phase, the criteria are divided into many components so it eases the evaluation (convert them to scores and scales to unify the language).
Independence: separate each interviewer from the others so they can make their own judgments.
Delay holistic judgments: after assembling all the data and facts, hires can make their final decisions by using their intuition.
As you can see, reducing noise has many benefits but as with any other thing, it has its costs too :
Dehumanizing/demoralizing: we often like to judge everything but what if we have no longer this ability, we will feel humiliated and respected and we can't use our morals neither to give a judgment.
Expensive: usually, we need to make our decisions as quick as possible, so sitting assessments, treating each one individually considering all information, and building algorithms free from noise require time and money.
Not feasible: some strategies are complex and you can't use them in certain situations more often in diagnostic cases(medical field).
Finally, we think noise is everywhere. Reducing it doesn't avoid all the mistakes but like hand washing prevents many of them.
Last thing but not least, to get the most benefit from these methods, we essentially need to compare the advantages and disadvantages that noise and bias can produce.