The book version of this article (The Checklist Manifesto) completely revolutionized how I did experimental science in graduate school. Early on, I saw many experiments ruined by human error: machine settings were incorrect, sensors not switched on, equipment not calibrated, incorrect instructions given to the subject, etc.
I was mortified by how little attention most professors and students paid to this problem. Something like half the data collections I was helping on (as a lowly first-year PhD) had some problem with the data collection.
I read The Checklist Manifesto and used it to design my own checklist -- and crucially, checklist procedure -- for my data collections and it made a huge difference. Not only did I collect much better data, but I was able to do much more sophisticated multi-step experiments without making mistakes.
The biggest takeaway for me was that checklists are not a piece of paper; they're a system. It's not enough for the surgeon to have a checklist, they'll just glance at it and think "oh of course I did these things". They only work when the nurse reads the checklist, item by item, aloud to the surgeon (I later realized this is exactly what SpaceX and NASA are doing during launch countdowns). Additionally, the checklist itself needs to be iterated on many times, removing unhelpful items and adding things you miss. And the checklist MUST be one page, no more.
If I ran an experimental lab of any kind, the book would be mandatory reading.
I'm a volunteer firefighter. One of the drawbacks of being a volunteer is that we don't typically get as much experience as our career counterparts, and that means in particular that we get less chance to develop muscle-memory and actual-memory of how to do things [0].
Consequently, I've come to have a much greater regard for simple checklists that make sure that we're doing basic stuff, we do it right. My favorite example: switching the fire engine's engine from driving the transmission to driving the pump. It's basically 3 or 4 steps, but when you only do this (at most) once a month (typically), remembering the steps is challenging, and it is even harder to remember to do the mandatory 2-3 second pause between each one.
The contrast with my life as a C++ developer for 35+ years is telling. In that world, almost everything is memory, and what isn't is both (a) look-up-able online (b) not time- or life-critical.
The brief time I've been a firefighter has made me very aware of a totally different side of my cognition - no, Paul, you actually don't remember most things at all - and has given me a great deal of respect for checklists to help prop up our minds in one of their weakest areas.
[0] truth is, I doubt that even the career firefighters in a lot of departments get enough daily experience for most things to become memorized, certainly not without many, many years responding to calls.
> The biggest takeaway for me was that checklists are not a piece of paper; they're a system.
I understood this when I read about plane pilots checklists. It's revolutionary even for everyday life. Every time we fail at something (for example leaving the vacation home with coffee still in the coffee machine), it goes to a checklist.
When something failed because it wasn't on a checklist, we say (wife and I) we lacked precision. And that thing now goes to the checklist.
Having to buy a new fridge because you arrive and something stayed in the fridge for x months ain't fun (not to mention you're left without any fridge for potentially two days [e.g. if you arrive on a late saturday evening]). It's on a checklist now.
Ever had to do a U-turn and lose one hour because wife forgot the phone at home? Checklist.
Weirdly enough checklists allows you to be more sloppy, not less. It's literally taking stuff out of your minds.
Ticket you receive at a toll? Checklist: "I gave you the ticket, do you confirm you received the ticket?" / "I confirm you gave me the ticket".
It's fun too. We all learned, including my kid, the NATO alphabet. It is precise.
We've got digital backups of our recurring checklists.
SpaceX and NASA countdowns aren’t quite checklists in the same sense of the surgeon’s or pilot’s checklist.
Those countdowns are verifying that a bunch of launch blocking gates have not been closed. The key difference is that it’s not a reminder for people of stuff that needs to be done. It’s more like a roll call to make sure they have a unanimous vote to go from all teams that have a stake.
I was mortified by how little attention most professors and students paid to this problem. Something like half the data collections I was helping on (as a lowly first-year PhD) had some problem with the data collection.
I read The Checklist Manifesto and used it to design my own checklist -- and crucially, checklist procedure -- for my data collections and it made a huge difference. Not only did I collect much better data, but I was able to do much more sophisticated multi-step experiments without making mistakes.
The biggest takeaway for me was that checklists are not a piece of paper; they're a system. It's not enough for the surgeon to have a checklist, they'll just glance at it and think "oh of course I did these things". They only work when the nurse reads the checklist, item by item, aloud to the surgeon (I later realized this is exactly what SpaceX and NASA are doing during launch countdowns). Additionally, the checklist itself needs to be iterated on many times, removing unhelpful items and adding things you miss. And the checklist MUST be one page, no more.
If I ran an experimental lab of any kind, the book would be mandatory reading.