0

A question states: Estimate the mean change in the number of flycatchers killed when the nest box tit occupancy increases by 10%.

The slope for this regression is .10766X. Hence, says the answer, take .10766(10) = the mean change.

My question is: The reason they are inserting 10 is not because it is a 10% change, correct? I cannot find another explanation but I am confident that this can't be the reason. Once I know for certain I can assess other possibilities.

user26091
  • 487
  • 5
  • 12
  • If `X` is a continuous variable in the regression, the coefficient is the change in `Y` for one unit of `X`. If you multiply the regression coefficient (slope) by 10, then this represents the change in `Y` for 10 units of `X`. – COOLSerdash May 29 '13 at 17:28
  • Right, but how does that relate to an increase of 10%? It doesn't, does it? It is not as though for example we are moving from 100 where 10% would be 10. Sorry if this is ambiguous but the question itself seemed rather ambiguous. – user26091 May 29 '13 at 17:37
  • A priori: no it doesn't relate to an increase of 10% in `X`. There is a special case: If `X` is log-transformed and `Y` not, then 1% increase in `X` is associated with a $\beta/100$ unit increase in `Y` (see [this post](http://stats.stackexchange.com/questions/18480/interpretation-of-log-transformed-predictor) for more information). – COOLSerdash May 29 '13 at 18:06

0 Answers0