15.5 No pooling

• Treat each runner separately, with a separate fit for each.

• Multiple fits, deal with each of the 36 runners seperately

Example for 3 runners:

• Seems great at first glance….

• However some significant drawbacks:

  • Cannot reliably generalize to groups outside our sample.

  • Assumes one group doesnt contain any information about another, ignores potentially useful information! (For example, do we really think that first runner gets faster with age?)