::
@timbo since we are only taking “successful candidates” wouldn’t that confirmed the survivorship bias that @mm12 raised?
As in candidates who only completed 1 or 2 levels are not taken into account, otherwise the average might be ‘increasing’.
The only way I can think of on how to achieve the 4 years average is as follows:
Starts Level 1 in Dec:
Dec 10 – passed
Jun 11 – failed
Jun 12 – passed
Jun 13 – failed
Jun 14 – passed
Total 3.5 + 0.5 (6 months revision) = 4
Starts Level 1 in Jun:
Jun 10
Jun 11 – failed
Jun 12
Jun 13 – failed
Jun 14
Total 4 + 0.5 (6 months revision) = 4.5
Average = (4.5 + 4) / 2 = 4.25
However, these are only my assumptions.