Checklists. It seems obvious that they work, but I almost never get a literal piece of paper with boxes to check, even where doing so would be simple and cheap, especially compared to the potential costs of skipping some step and doing it wrong.
There are many ways to do checklists wrong, though. They can easily lead to bureaucracy is not applied correctly. Maybe that’s why many people got burned by them. Checklists must be short.
Agreed, I read at least part of The Checklist Manifesto, and they are super helpful and underrated. Even for processes which seem most amenable to a checklist (clear steps that don’t change much), a checklist is still often missing.
Full agree. I use checklists all the time in research, and there is a big difference between an actual checklist and an almost-a-checklist research protocol. They are easy to make, available to anybody right now, and way neglected.
A lot of protocols are written in recipe format: a list of steps, and maybe a list of reagents. A lot of individual steps are actually compound, and there are extra steps that are useful and important but not written (ie check if you’re low on reagent X and order more if so).
It’s easy to lose your place in the protocol (did I add buffer Y already?), to skip implicit steps (did I already mix the buffer in?), or to leave things out (ie the middle bit in a single instruction). You can add check boxes to such a protocol, but it’s easy not to think of that.
Checklists. It seems obvious that they work, but I almost never get a literal piece of paper with boxes to check, even where doing so would be simple and cheap, especially compared to the potential costs of skipping some step and doing it wrong.
There are many ways to do checklists wrong, though. They can easily lead to bureaucracy is not applied correctly. Maybe that’s why many people got burned by them. Checklists must be short.
Agreed, I read at least part of The Checklist Manifesto, and they are super helpful and underrated. Even for processes which seem most amenable to a checklist (clear steps that don’t change much), a checklist is still often missing.
Full agree. I use checklists all the time in research, and there is a big difference between an actual checklist and an almost-a-checklist research protocol. They are easy to make, available to anybody right now, and way neglected.
I’m interested in hearing more about almost-a-checklist protocols and what’s up with them.
A lot of protocols are written in recipe format: a list of steps, and maybe a list of reagents. A lot of individual steps are actually compound, and there are extra steps that are useful and important but not written (ie check if you’re low on reagent X and order more if so).
It’s easy to lose your place in the protocol (did I add buffer Y already?), to skip implicit steps (did I already mix the buffer in?), or to leave things out (ie the middle bit in a single instruction). You can add check boxes to such a protocol, but it’s easy not to think of that.