Reply
Tue 12 Apr, 2011 10:48 pm
If I have a residual graph and say, there is a region/cluster of points in my graph where more points are above the x axis than below (eg. 25 points above x axis, 12 points below the x axis. x value is between 25-60), would the residual graph still have a random pattern/distribution and would a linear model still be a suitable fit? There does not appear to be any curving.
@qwertyqwert,
Yes, I think it's a random pattern. Why would you think otherwise?