Availability Error

Availability Error

David Kyle Johnson

I’m never getting in a plane again. They’re just too dangerous. Seems like there’s another plane crash every month. From now on, I’m driving everywhere I go. Anonymous

One commits the availability error when one pays attention to, or is compelled by, the readily available evidence – the evidence that is obvious, memorable, or psychologically compelling – instead of taking into account all the evidence or the reliable evidence. If you believe that a full moon makes people crazy (the so called “lunar effect”), and take special notice of crimes committed during full moons while ignoring the same kind of crimes on all other nights, you are subject to the availability error.

The availability error contributes to confirmation bias, the tendency to only pay attention to the evidence that confirms what you believe and ignore the evidence that doesn’t. But the error can also just appear on its own. In the quotation that begins this chapter, the person (who shall remain anonymous) is only paying attention to the stories about plane crashes she hears on the news and ignoring the statistical evidence that traveling by automobile is far more dangerous than traveling by plane. The source gives historical lifetime estimates of 1/98 for dying in a vehicle accident and 1/7178 for dying in a plane crash. These are population estimates with different travel exposure, rather than probabilities for a particular journey. Why do people think plane travel is less safe? Because plane accidents are so rare and compelling, they get news coverage and thus are immediately apparent to one’s mind. Traffic accidents, on the other hand, are so common we don’t even think twice about them. A comparison of particular journeys needs risk estimates measured on a comparable basis.

The availability error also encourages other fallacies, such as hasty generalization – where you generalize about an entire group based on an unrepresentative sample of that group. If you conclude that all members of some race are violent criminals just because you noticed a few people of that race that are, you are paying attention to the psychologically compelling evidence (the stories about the criminals) and ignoring the better, less compelling evidence (the much larger number of people in the group who are not violent criminals).

Many people try to take advantage of our propensity to commit the availability error in order to fool us. Self‐professed psychics, mediums, and astrologers are great examples. Psychics take advantage of this by putting out hundreds of predictions at the beginning of the year and then only mentioning the ones they got right at the end of the year. Or consider this real‐life dialogue from an astrological reading given on one of my favorite television shows, Penn & Teller’s Bullshit:

READER: Are you living near a railroad, or a diamond store, or a government building? SITTER: There’s a train track in front of my complex now. READER: Yeah, that’s the Capricorn Moon.

The reader is actually using two tricks here. One is that he is stating things that are true of almost everyone; railroads, diamond (jewelry) stores, and government buildings are everywhere, so most people live near at least one. (I currently live near all three.) But he is also counting on the sitter’s focusing on what he gets right (in this case the train track) and ignoring what he doesn’t. In another part of the reading, the reader asked this same sitter whether she lived near a jungle, the woods, the country, a zoo, a church, a courthouse, a publishing house, a college, or the outdoors. She answered no to all, yet afterwards she still ranked the accuracy of the astrologer’s reading as a 10 out of 10. Why? Because of the availability error. She’s remembering the hits and forgetting the misses.

In fact, memory is especially susceptible to the availability error because we are apt to remember only what is memorable, and that is often only what is psychologically compelling. This is why people often think they always get stuck in the slow lane in traffic jams. Think about it. When you are stuck in the slow lane, what’s it like? You’re frustrated, it’s likely to make you late, you’re there for a long time, and you’ll stew about it for hours. But when you’re in the fast lane, you’re calm, you’re not there for long, and you’ll likely forget about it a mile later. Memorable delays can dominate your impression. To establish how often each occurred, you would need a record of both, rather than assume equal time in fast and slow lanes.

Or take another experience that most of us have had: being suddenly reminded of a dream you had because of some event that happened during the day. Such an event causes some people to suspect that they have psychic powers – that they can dream the future before it happens. But in reality, we should expect for such a thing to happen to us every now and again. According to Ted Schick and Lewis Vaughn (2014), authors of How To Think about Weird Things, although we only experience four or five periods of REM (rapid eye movement) cycle sleep a night, we normally experience at least 250 dream themes each night (117). We of course forget most of them and thus ignore all the counter‐examples that show that our dreams don’t predict the future. So when an event during the day triggers a memory of something we dreamed, it may seem remarkable – but a remembered resemblance can occur by chance.

Take the story (mentioned in chapter one of Weird Things) of someone who dreamed of not being able to save his father from falling off a cliff only to (in real life, three weeks later) fail to save his father from falling out of a second‐story window to his death while painting a windowsill. This too may seem remarkable but (a) this person is failing to recall the hundreds of his dreams that didn’t come true and (b) we are failing to recognize all the people to whom this hasn’t happened.

Many dreams, events and possible resemblances create opportunities for apparent “premonitions”. Counting only the matches leaves out the dreams that did not correspond to later events. A resemblance alone does not establish supernatural foresight.

But this is not the only way that memory can lead us astray. In fact, memory is notoriously unreliable – much less reliable than people assume. We think of memory as being like a video camera, able to record events and play them back just as they happened (see Maynes 2011). In reality, the process of storing and recalling memories is highly unreliable and apt to failure. When we store a memory of an event, we don’t record all the specifics. We store general themes and leave out the details. When we recall that event, we harken back to the theme but fill in the details according to what might make sense – or even to what would make the memory align with our expectations. When we store the memory away again, some of those confabulated details can be stored away with it. A later recollection can draw on earlier reconstructions as well as the original experience (see McDermott and Roedifer n.d.). Consequently, those confabulated details can seem just as accurate as things that actually happened.

Recollection creates an opportunity for details to change, rather than inevitably making a memory less accurate each time (see Fernyhough 2014). This is why, when you go back and watch a movie you only saw once but haven’t seen in a while, individual scenes (or even major plots points) are often very different than you remembered them being the first time.

Research discussed here examines the limits of eyewitness memory (see Arkowitz and Lileinfeld 2010). An eyewitness’s confidence alone does not establish that a recollection is accurate (see Crowell 2011). We all think we can justifiably believe that something happened if we saw it with our own eyes, yet the research suggests that far more often than we realize, our memory of what we have seen with our own eyes is not accurate.

Even memories of large events – so called “flashbulb memories,” like your memory of where you were when you heard about 9/11 – aren’t as reliable as we assume. Now they do seem especially clear and vivid, and we do recall them with confidence. But in one study, led by New York University psychologist W. Hirst (2009), researchers asked and recorded the details of where people were, what they were doing, and whom they were with when they first heard of the 9/11 attacks. The passage reports their memories as 63% accurate at a follow-up one year later and 57% after three years.

This wasn’t a fluke. Many such studies have been done and they all show basically the same thing. To be fair, it seems that most of the confabulating and forgetting that will be done has been done after three years. The flashbulb memories of 9/11 weren’t that much less accurate 10 years out (see Law 2011). But still, 57% is much less accurate than most assume our flashbulb memories to be. In fact, that’s probably less reliable than you assumed your normal memories to be.

Evidence that feels “psychologically compelling” or “most memorable” may be unrepresentative. Before judging frequency or likelihood, compare what comes to mind with a record that also includes the less memorable cases.