Is your memory damaging your judgment?
We all draw on our memories to help us judge situations and create solutions, yet sometimes our memories cause us to come to the wrong conclusions. At fault may be the Availability Heuristic or bias.
Preface: This is not intended to be a political blog, but I may use political examples. They are examples and may or may not be what I believe.
I noticed a fascinating example of the Availability Heuristic towards the end of last year. As the debate on Mask Mandate or No-Mask Mandate started, many people responded directly based on their experience. Without any reference to data, people's response to the need for the mandate seemed directly related to if they or someone they knew had had COVID. If they or a family member had suffered, they would often make emotional cases for everyone wearing masks.
Conversely, those who did not know anyone who had been touched by COVID tended to believe that COVID was not as significant. Should these two groups meet on Twitter, you would find a vigorous debate, still with no data, about how serious a problem COVID was. To me, this was a great example of the Availability Heuristic. Of course many came to the debate with what they thought was the ‘right’ data, but mostly it was data that proved their point – see Confirmation Bias.
Your data may confuse you
The concept behind the availability heuristic is that if you remember something, it is probably more crucial than something you can't recall. The implication is that we tend to give more credence toward more recently received information and biases than toward the latest news.
After approaching nearly two years of wall-to-wall coverage of COVID and its potential dangers, most Americans have an overinflated view of how dangerous it is. As Bill Maher, not a noted right-of-center commentator, pointed out on the Jimmy Kimmel Live! Show:
MAHER: Now the reason why this is relevant, I have to cite a survey that was in the "New York Times," which is a liberal paper, so they weren't looking for this answer. But they were talking about -- the question was, what do you think the chances are that you would have to go to the hospital if you got Covid? And Democrats thought that was way higher than Republicans. 41% of Democrats -- the answer is between 1% and 5%.
KIMMEL: Okay.
MAHER: 40% of Democrats thought it was over 50%. Another 28% thought it was 20% to 49%. 70% of Democrats thought it was way, way, way higher than it really was.
So, my point here is not whether Democrats or Republicans have it right, or even if COVID-19 is or isn't dangerous. Intelligent people thought the chances of being hospitalized were over seven times more than it was.
Why?
Well, partly because we watch the news, and the news reports death and hospitalizations like sports scores without real context. Why they do is another exciting conversation for another day, the fact that they do creates the effect that something is more prevalent. Our Availability Bias means we overestimate the likelihood of being hospitalized because of all the data we get. Interestingly, the more emotional or unusual the information is, the greater impact the information has and the more relevant it may appear.
So how do make sure availability bias has a limited effect on us?
Frequency and probability
In the original research behind this bias, Tversky, and Kahneman (see Thinking Fast and Slow) showed three significant factors that may cause this. They cited them as the frequency of repetition, frequency of co-occurrence, and illusory correlation. The idea here is that the more an instance is repeated within a category or list, the stronger the link between the two instances becomes. They say COVID and HOSPITALIZATION enough, and you start to connect the two beyond the actual correlation.
So, how do you break out of this bias?
Here are 3 steps that might help:
1. Always understand the data
In the middle of 2021, I saw a conversation about the percentage of people dying from COVID-19 who were not vaccinated. The person who raised the data said that well over “90% of the people who were dying from COVID had not been vaccinated.” The problem was that the count for this data was from January when no one was vaccinated.
Again, I am not saying you should or should not be vaccinated (ask your doctor); I observed that the person using the data was trying to imply that data was about ‘that day and not over a more extended period.’ Check the information, ask for the sample, check for base rates … never believe (or repeat) data without knowing where it came from.
2. Regression to the mean and trends
Do not look for data; look for trends because outlying data can mislead and often is not repeated. This notion, called regression to the mean, was first worked out by Sir Francis Galton. The rule goes that, in any series with complex phenomena that are dependent on many variables, where chance is involved, extreme outcomes tend to be followed by more moderate ones.
3. Question your assumptions
Whenever you draw a conclusion from data, take a moment to question your assumptions and ask yourself some questions:
Is this what I expected to hear?
Does it fit too closely with my beliefs?
What data can I find that might prove the opposite?
If nothing else, count to 10 before believing, and give yourself time to be skeptical of what you are hearing.
Check your biases at the door
Heuristics or biases are impactful because they are shortcuts to get us through our daily lives. Without these shortcuts, life would be exhausting and unmanageable. But like all shortcuts, sometimes they make life easier, and sometimes they lead you astray.
Once you know about Availability Heuristic or Bias, you can look out for it. You may also want to go back to some of the things you have said or believed recently and see if they still hold up. Some of what you may believe may be stories you are telling yourself.