Why Recent Memories Make Terrible Advisors
Is that really true or is it just the last thing you remember?
It was freezing that morning. I put on my helmet, my bike lights, my gloves. In a flurry, I left the main door, carrying my bike with me, placed it on the road and off I went. It felt no different than any other morning. I was cruising through, cold wind piercing through my face that I secretly enjoy, makes me feel alive.
I arrived at a mandatory crossing and this is my usual route so I am well aware that I will stop here. I break and put my boots on the ground and suddenly, my boots slip, unable to find any grip, unable to stand up, with a thud I land on the metal frame (I don’t have a “ladies frame” bike) Ouch. I now realize the ground beneath me is ice. I am straight out of a scene from a funny cartoon, my feet slipping, wobbling to find balance.
It was not straightforward to stay upright during this next step but somehow I am able to swing my right leg off my bike and inch-by-inch cross that road on foot.
“Let’s try again” I said to myself and start riding again, cautious. A few meters further, out of the blue, my wheels starts wobbling in all directions and once again, I fall, to the ground this time, very gently but a fall nonetheless.
“Are you serious, bike!” I said to my bike.
That was enough adventure for one morning. I was not going to test the “Third time’s the charm”
A couple of days later, I took out my bike again. Cautious as it was still icy, I was hesitant to step on the ground because the recent slippery events were still top of mind. This time, all went well. Next time, I only thought of my icy fall half way into the ride and soon enough, the fear melted away somewhere into the back of my mind, as did the snow.
Now if you had asked me about the odds of falling in icy/snowy roads ONE DAY before my fall, my answer would have been some random number and if you had asked the same question, A DAY later - my answer would have been suddenly a lot higher. The real rate of falling is probably not changing with my one fall and probably does not change a lot year on year (it must freeze as a given).
This means my own recent experience of falling is influencing my answer to how likely it is. As normal as that is, it leads to wrong answers because that is, simply not how facts and maths work.
This reveals a bias in quick thinking: the recency bias.
Recency bias is the cognitive tendency to give disproportionate weight to recent events, experiences, or information when making judgments or decisions, while undervaluing older but potentially more relevant data.
This was a low stakes example, right. But Recency bias is prevalent, it shows up everywhere.
When the stakes are low
⛷️ Holiday destination
You tell your friends and family about where you’ll go on holiday next. “Ah” they say, “I was there last year and the weather was shit” — Well meaning but they’ve just enacted recency bias. They remember their last data point and unfortunately, they got rain. But does it actually do justice to the weather you should expect? Well if the chances of raining were 2% .. guess what, it can STILL rain.
👎🏼 Bad experience
You went to a new restaurant with a top rating and the food/service was not great. Guess what? It’s highly unlikely that you’ll be going there again because even though it may have been a coincidence, the emotional cost or the opportunity cost of going to that restaurant was high enough (for most people who don’t go to restaurants every second day) that you would not risk it again. I’m not judging your choice — I would do the same but it may not be rational.
When the stakes are high
🤨 Too many foreigners?
Imagine you walk out into a street in your country and you see a big group of foreigners. Right after, you are interviewed “Do you think there are too many foreigners in this country?”
💡 Did you know that you are most likely to base your answer NOT on your knowledge of truth (how many foreigners as % of population and your assessment of that) — you are instead very likely to base your answer on this recent memory. This can lead to a really skewed view of reality that’s hard to change.
🤺 Fighting over not emptying the dishwasher?
Lot of marital arguments happen over expectations and disappointments and a mismatch over reality. Studies consistently find that majority of partners believe that they do MORE THAN THEIR fair share of chores. Let’s say a woman thinks she does 80% of the chores and her partner thinks she does 50% of the chores — well there is clearly a problem here because together we often end up with a combined reporting of more than 100% .. which is factually impossible.
💡 The reason why partners overestimate and eventually argue about the distribution of workload is because they are heavily influenced by recent events and only their own share of events. Each partner vividly remembers every time they emptied the dishwasher, but their partner’s contributions fade into background noise. She remembers her own three loads yesterday; her wife’s blur together over time.
So now you’re thinking: Well, is this a problem?
Is recency bias bad for you?
Yes and no.
Most biases exist to protect us and are useful on a day to day basis. Recency bias evolved because what’s RECENT used to be more RELEVANT.
Just saw a grizzly bear down that river? Well that’s relevant and recent.
Saw a grizzly bear down 3 months ago? Useless.
So our brains have evolved to give more emphasis to recent events because they usually represented highly-valuable life-or-death information. Closely related is our tendency to avoid uncertainty. It goes hand in hand. Recent information is marked important to our brains and it will be used to make a decision quickly because not making a decision means sitting with uncertainty and humans dislike uncertainty, very much. For the very same reason, uncertainty that pesters too long means you’ve now become grizzly food, too bad.
🤖 A buggy feature
Is recent information always more relevant? No, not in our fast-changing, modern, hyper connected world where historical events do matter and do influence the results.
We’ve been left impressionable and prone to making harmful decisions. And impressions are hard to change. Once we land on a conclusion, an opinion, its really no easy feat to change our minds.
Today, recency bias nefariously contributes to
hate and violence: one case of a woman abusing the alimony system during her divorce to extract money is circulated repeatedly on social media. the same content recycled and shown and used to create outrage and hate. When people are then asked if they would vote to keep laws protecting the weaker party in place? Many will say no. The algorithm shows you the same outrage-inducing case repeatedly, creating an illusion of frequency. Your brain mistakes recent exposure for statistical prevalence.
missed opportunities: You’re making an important decision choosing between two cities for your new job. Your friend had a really bad job experience in one of the cities. Do you know this story that comes top of mind (because of it’s emotional hold) can sway your decision? In practice, it should not influence you but in reality, it will.
discrimination: you hired someone from a specific nationality and they turned out to be very shitty at their job. When re-hiring, you see a candidate with the same nationality and your brain gives a subtle signal “nope, pass” — this is recency bias in action. Recent bad experience comes to mind and signals danger. By listening to this signal, you strengthen this now subtle belief that “those people are really bad at this job”
and just plain bad judgement: people make wrong decisions all the time simply because they prioritize recent events over long term data. Be it bad investment — extrapolating recent market trends into the future. Be it weird superstitions in sport like “the hot hand” — no, just because he scored a goal 3 times in a row, he does not have a higher chance of scoring the next one. Evaluating people wrong — doing bad in the last month led to a bad evaluation despite doing GREAT for the 9 months before because her boss got swayed by recency bias.
👑 One rule to rule them all
So how do we avoid the harms of Recency bias? specially when we know the downsides can be huge.
There is one rule to rule them all, and it’s gold:
When the cost of failure is low, decide quickly. When the cost of failure is high, slow down.
It speaks for itself, it’s elegant and it works.
Biases are these automatic behaviors that are not BAD in and of themselves. Intelligence is when we can develop the muscle to quickly assess when do we want to decide quickly (and let the biases do their work) and when it’s time to take it slow (and engage logical, deeper reasoning)
A very good parameter to judge: the cost of failure. What is at stake if you make a wrong decision?
Go back to the previous examples and you’ll see that sometimes the cost is low — a lunch, a holiday destination but sometimes its high — hate crimes, racism, marital rifts and financial losses
I’ll end with what Daniel Kahneman said (paraphrased)
When deciding, break the problem into multiple dimensions and really think about each dimension separately and independently. Don’t allow people to give the final judgement too early. If you find a decisive dimension, you stop and if not, keep going. Your decision will almost certainly be better.
🔄 TL;DR
Your mind marks recent information as more important than older experiences and historical data, it has most likely evolved to do so
This causes recency bias where, in decision-making, you will have a tendency to over-index your decisions on “what easily comes to mind” instead of diving deep into what the correct or true answer really is.
This tendency is OK 95% of the times because most decisions are low-stakes.
This tendency will lead to poorer decision making in high-stakes decisions.
Thus, when the cost of failure is high, go very very slow and deliberate on the decision as analytically and independently as possible.
Thanks for reading and because you made it until the end, please enjoy this butterfly art 🦋
