[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: A vvvvvery naive question about "controll for"
Dear Statalist users,
I wish I'm not asking something very naive. Does including a var as an
additional regressor mean this var is being controlled for? In
physical science as far as I understand, "control" means maintaining
some conditions at the same level for all the subjects under
experiment; but in social science this is hardly true, so putting one
variable indicating the difference of one aspect among these subjects
really "controls" this aspect? If this is true, then I come to an
absurd conclusion that physicists should learn from social scientists:
putting more and more variables into one regression instead of
spending lots of funds in building up expensive labotories to
"control" some things.
I doubt so but never figured out a satisfactory answer. Any opinions?
* For searches and help try: