The 28:1 Grant/Sackman legend is misleading, or: How large is interpersonal variation really?

How long do different programmers take to solve the same task?
In 1967, Grant and Sackman published their now famous number of
28:1
interpersonal performance differences, which is both incorrect
and
misleading.
This report presents the analysis of a much larger dataset of
software
engineering work time data with respect to the
same question. It corrects the false 28:1 value, proposes more
appropriate
metrics, presents the results for the larger dataset,
and presents results of several further analyses:
distribution shapes, effect sizes,
and the performance of various significance tests.