nyashfan
For Earth Below
Shouright - you must have a PhD in statistics! Now we are talking about correlations pertaining to conditional bivariate normal distributions...
Bottom line is that rushing attempts are leading to 1st downs by rushing. It's not just a matter of compiling a bunch of rushing attempts. There is a subtle causal relationship in which rushing attempts lead to first downs which leads to controlling the clock, fewer offensive turnovers, better down and distance situations, and so forth. And as someone else said how important it is to amass rushing 1st downs/attempts in the 4th quarter in order to close out games.
Bottom line is that rushing attempts are leading to 1st downs by rushing. It's not just a matter of compiling a bunch of rushing attempts. There is a subtle causal relationship in which rushing attempts lead to first downs which leads to controlling the clock, fewer offensive turnovers, better down and distance situations, and so forth. And as someone else said how important it is to amass rushing 1st downs/attempts in the 4th quarter in order to close out games.
Indeed. The correlation between rushing first downs differential and rushing attempts differential is .90.
However, the partial correlations between winning percentage and both rushing first downs differential and rushing attempts differential drop to a level of meaninglessness when one or the other of the latter two variables is controlled for, meaning that we likely have a circular sort of relationship between them as they relate to winning percentage.
One has to attempt a rush, however, before one can get a first down on one, which likely puts rushing attempts in a causal position here.
http://en.wikipedia.org/wiki/Partial_correlation