This course will become read-only in the near future. Tell us at community.p2pu.org if that is a problem.

Statistical Blips [Sept. 12, 2012, 6:13 a.m.]


 

Dilbert.com

 

Task: Yes, this sounds very strange. But please bear with us. Grab a piece of paper and write the letters A to J. These are the names of ten individuals who work for you.

You need to toss a coin six times for each individual and count how many heads they get. Write the number down. 

For example when I do this, I get the following: 

Name                  Number of Heads
Albert 3
Bertie 5
Charlie 3
Dave 3
Eddy 2
Freddy 1
Gina 3
Hannah 6
Ian 4
Joe 3

 

Dilbert.com

You now need to sack the incompetent workers who only got 0, 1 or 2 heads. You should promote your best performing employee. Try again (there's less coins to throw this time).

 

Please do try this exercise for at least one round, stop, and think about it. Discuss it with a colleague. Think about it again.

I'm now going to suggest that what we are doing is rewarding and punishing people for randomness. On average you get three heads, but ocassionally you get zero, ocassionaly you get six. Quite often you get 2 or 4. Now, we know that the chance of coming up heads is exactly 50:50. But other than that, how is this different from a performance indicator? What if you were expected to keep the number of fires that spread beyond room of origin to less than 1 per 50,000 households. Is this a bit like a coin that is 1:50,000 rather than 50:50. Why should you instantly punish someone who gets 2 per 50,000 rather than 0 or 1?

Please do think very carefully. Please indicate if you agree with this suggestion or if you think there are difficulties with this approach.