Multiple Regression R^2 Issue
When your model has as many free parameters as data points, then there are
zero degrees of freedom for lack of fit. The fit will be perfect and R^2
will be one.
Jerry
"DigitalDuck" wrote:
However, when I run a single regression on just one variable's data, even
though it is only 3 data entries, it gives me a R^2 anywhere from .08 to .98
which would make sense to me. Why when I add multiple variables would it
then change to being 1 when it seemingly works for single variable?
"DigitalDuck" wrote:
That makes sense. So in order to have it not = 1 I need to add more data
points for each variable, correct?
"Jerry W. Lewis" wrote:
I am not sure what you were intending to do, but you appear to have asked
Excel to fit 3 variables (or 2 variables and an intercept) to 3 data points,
which by definition will fit perfectly (though is unlikely to be meaningful).
Jerry
"DigitalDuck" wrote:
Whenever I run a multiple regression it always tells me that my R^2 is 1
Now I know these numbers aren't that good.
Maybe someone can help me figure this out.
Here is what it outputs in the summary field:
Regression Statistics:
Multiple R: 1
R Squa 1
Adj. R Squa -1.2E-09
Standard Error: 1.15E-10
Observations 3
I have 7 input variables each with 3 observations.
Anyone know what is going on or if I am just screwing it up and reading it
wrong?
I haven't been in stat for a while, and when I was I used SPSS which was way
more user friendly.
Thanks.
I know there are very ef
|