During World War 2, a statistician named Abraham Wald was asked by the US Navy to recommend ways to minimize bomber losses to enemy fire. As a starting point, he was provided with data about where bombers that returned to base had been hit. His data looked a bit like this:
Looking at this data, Abraham recommended to the Navy to reinforce the armor in the parts of the planes that hadn’t been hit.
Wait, what?
This is counter-intuitive, but remember where his data came from: planes that made it back to base. He assumed (correctly) that the parts of the planes that didn’t have bullet holes were the parts that, if hit, would destroy the plane and kill the crew.
This is a real-world example of survivorship bias. Survivorship bias leads us to conclude all sorts of wrong-headed things about the world. Here are two examples you’ve probably run into:
“They don’t make ‘em like they used to.” Things made in the past are supposedly better than the obviously cheap, low-quality things of the present. In reality, most things in the past were low quality too, and broke or were discarded such that only the best stuff remains today.
“The best entrepreneurs are college dropouts.” Bill Gates and Mark Zuckerberg are enormously successful and both of them dropped out of college. Dropping out must, therefore, be a good way to achieve business success! This ignores the fact that the overwhelming majority of businesses fail and college dropouts earn less on average than graduates.
In both of these cases, the selection pressure (things breaking over time and people failing at business ventures) are ignored. This makes backward causal relationships look plausible. Survivorship bias in action!
The Sampling Problem and Career Advice
If you’re trying to get ahead in your career, whose advice would you pay more attention to?
A coworker who has 15 more years of experience in your discipline, but is the same level as you in your current organization.
A Vice President in your discipline at the industry-leading company with 10 fewer years of experience.
The overwhelming majority of people would choose to listen to the younger VP. The justification for that preference is touted as simple common sense. How can you learn the best path to get promoted to senior leadership by someone who hasn’t themselves accomplished that feat? Plus, the second person must be extra talented because she achieved a senior title while still young!
But that’s backwards. It’s akin to looking at the airplane damage diagram above and concluding that the way to save planes is to add armor to the wingtips.
Most conventional career advice demonstrates both extreme sampling and survivorship bias.
So What Can You Do?
So how should you acquire actionable, unbiased career advice?
Simple: solicit advice from many people who are a little ahead of you that are also demographically similar. Let’s break that down a bit:
Getting Lots of Advice
Everyone’s path is different. What works for one person won’t work for another. Maybe you need to stay in a particular town to take care of an aging relative. Maybe your top priority in life is to travel the world, and staying in one place is a non-starter.
Everyone works within different sets of constraints. If you rely on only a few career mentors, you’ll get advice that isn’t helpful to your circumstances.
A Little Bit Ahead of You
When you are learning a skill, it’s tempting to seek out a teacher that is the best of the best. But that’s actually counter-productive.
Someone who has 30 years of experience playing the violin suffers from a profound curse of knowledge: they’ve forgotten what it’s like to be a beginner.
You would be better off finding a teacher who is only a couple of years further along than you. They are more likely to remember what it’s like to learn basic posture and finger positions. They can immediately recall the experience of not knowing what they’re doing and help another person with that transition.
Demographically Similar
The world is deeply unequal and unfair. The sooner you understand this, the more effective you’ll be at furthering your career.
If you are a middle-aged mother of 3, the way that you succeed is quite different from a 23-year-old recent grad and first generation US immigrant. Society has all sorts of ‘isms that constrain both people, but they are radically different. People that don’t experience the world in a similar way are likely to give you bad advice.
Putting It Together
Hopefully in this post, I’ve convinced you that most career advice is essentially unusable due to survivorship and sampling bias.
Luckily, it’s not the end of the world. By averaging together the suggestions of lots of people who are a little ahead and similar to you, you can accelerate your career better than any Master Class could ever do.