The date you changed something, and the ordinary days
When you change a condition of your own work, a new hour, a new desk, a break you did not take before, write down the date and what you changed. On the other days, write down that nothing changed. The page shows the change and the days on either side of it, and it never tells you whether the change worked.
Today
Keep the days on which nothing changed. Without a record of the ordinary days, the changed day will always look like a cause.
The last six weeks
no linenothing changeda change
Every line
No averages before and after, and no chart of how the days went. The record shows the change and the days on either side of it, and stops there.
What the record says. The one practice in the relay room story that holds up is the log: a record, kept at the time, of what was done and on which day. It is the dated record that let Levitt and List find the Mondays: in the lighting tests all the changes were made on Mondays, and output rose just as much on Mondays when nothing was changed at all. If you change something about the conditions of your own work, write down the date you changed it and what you changed. That line costs nothing and it is the only thing in the relay room that anyone has ever been able to check. Your record will show you the change and the days on either side of it. It will not tell you whether the change did anything, and neither did the most famous record of this kind ever kept. The idea that people work better because they are being watched was named after this room a quarter of a century later, by people who had not been there, and in 1992 Stephen Jones found "slender or no evidence of a Hawthorne effect" in the room's own data. What is left is small and clerical, and it is enough: a date, a change, and the honesty not to claim what the record cannot show.