This year we purchased an automated communication system that allows our district to readily contact all or selected families by phone, email, or text message. With this technology available we have the opportunity to contact all families in the event of school cancellation due to snow/weather conditions. Before moving ahead with what I thought (without any data) was a great idea, I decided to get input from those most impacted: Families. It appeared that the “loss function” process would be the best way to get the input from families I desired, and to this end I created an electronic survey with one main question:
Are you in favor of being notified by phone when school cancellation takes place due to snow/inclement weather?
If the answer to this question was “no,” the survey ended. If the answer was yes, it led to three more questions (this is the loss function part of the survey), which I refer to as the “Goldilocks paradox:”
What time would be too early to call?
What time would be too late to call?
What time would you prefer to be called?
The possible choices to answer these three questions included:
5:00 a.m.
5:15 a.m.
5:30 a.m.
5:45 a.m.
6:00 a.m.
6:15 a.m.
6:30 a.m.
6:45 a.m.
7:00 a.m.
I received 683 responses. 88% were in favor of using this communication method and 12% were not in favor. The data clearly informed me to move ahead. The loss function process showed the least number of “losses” at 6:15 a.m., while the two most “popular” “best times to call” were 6:00 a.m. and 6:30 a.m.
Since 6:15 a.m. had the least “losses” and since it was wedged in between the two most favorable times, the data has informed me the best time to make the automated call for school closures is 6:15 a.m. We will also communicate our school closing on our local media outlets and our website.
The loss function process was extremely helpful in the decision making process, for the only other way I was going to make a decision (and I did consider this) was just to make my “best guess” using my “gut feeling.” I believe the outcome of the process has produced the best time to call and the decision can be defended to anyone based on the data that was collected.