Biased against bias? Why we dislike bias, and why we often need it.

Disclaimer: some of this article was written using the assistance of ChatGPT-4, a large language model (LLM) produced by OpenAI. Portions of this article written in part by ChatGPT are clearly marked by section dividers. Text written by ChatGPT has been modified to fit the purpose of this blog.

Last week we talked out heuristics, or mental shortcuts, which make decision-making faster. This speed, of course, comes at a cost. The use of heuristics (most commonly in intuitive thinking) are often biased; they systematically favor a particular conclusion more than they should. This makes biased decisions both less fair and less accurate. So, it sounds like an ideal world for decisions is one in which bias doesn't exist, right? Not exactly.

It's important to first note before continuing that when I say "bias," what I do not mean is social discrimination. Bias in the context I use it refers to methods of gathering or analyzing information which favor a particular conclusion. For example, a man who routinely underestimates how much money he will spend at the grocery store biased. A judge that routinely sentences black defendants to longer prison sentences is biased as well as socially discriminatory. Not all biases result in social discrimination, but all social discrimination results from bias. A world without social discrimination is ideal, though a world without any kind of bias is not, as you will read below.

Begin ChatGPT-4 writing

While it's true that bias can lead to inaccurate conclusions and unfair decisions, it's important to recognize that not all biases are inherently negative. Generally, biases serve as adaptive mechanisms that help us navigate the complexities of the world. They accomplish this by arriving at workable conclusions given only limited information and time. To understand the potential benefits of bias, let's examine a couple examples of how it can actually be useful.

1. The Availability Heuristic / Biased Recall

Perhaps the most ubiquitous heuristic of all is the availability heuristic, a mental shortcut wherein a person relies only on information readily available, namely information in the person's memory. The problems with this approach will be further illustrated in a later article, but suffice to say for now that a person's memory is rarely representative of reality (representativeness problem), and our recall from memory is even less so (biased recall). The upshot is that by relying on memory the search for new information can be entirely skipped.

For example, imagine you're in an unfamiliar city and need to choose a restaurant for dinner. You might rely on the availability heuristic by selecting a place you've seen advertised or heard about from friends. In this case, the bias helps you make a decision without spending excessive time researching every possible option.

2. Positivity Bias

Positivity bias is the tendency to seek evidence which will confirm our hypotheses rather than disconfirm (falsify) them. This approach has its advantages, particularly in situations where time and resources are limited. The positive testing strategy allows us to narrow down our focus and gather relevant information more efficiently.

For example, suppose you're trying to find a plant for your garden which is safe for your pets. A positive testing strategy might involve researching plants known to be pet-friendly and listing them as potential candidates for your garden. By focusing on positive outcomes (i.e. safe plants), you can quickly establish a solid understanding of which plants pose no threat to your pets.

Bias vs. Objectivity

Given the complexities of our environment and the limitations of our cognitive processing abilities, biases can serve as useful tools for navigating uncertainty and making decisions in a timely manner. The focus of the study of bias on its shortcomings is reasonable insofar as they demonstrate situations in which we should learn to recognize our biases and address them. After all, if a bias were always to lead to the best possible conclusion, then there would be no need to address it, as doing so would afford no benefit. Unfortunately, this focus on the shortcomings of bias has led to a popular perception that biases are inherently bad. I'm hoping this article has helped show how this is not the case. Future articles on bias will focus on those situations where bias should be recognized and avoided, but never will I say that they are inherently bad.

Up next: What do bomber planes teach us about vaccines?