# F test and categorical variables

I am trying to compute a regression.

I have a continuous dependent variable (investment in millions of dollars)
One continuous independent variables
Five dichotomous indicators (1 / 0)
Four interactions.
Robust standard errors are being used.

N=569.

Is there any particular reason why the F-Test is not being worked out in Stata?

Stata help says:

Your estimation results show an F or chi2 model statistic reported to be missing. Stata has done that so as to not be misleading, not because there is something necessarily wrong with your model

The problem seems to be related to the interactive dummy variables, because when I remove them the F test is calculated.

The regression output is the following

Linear regression                                        Number of obs =  580
F(  9,   569) =  .
Prob > F      =  .
R-squared     =  0.2336
Root MSE      =  1.2608

Robust
lninv                         Coef.       Std. Err.  t      P>t    [95% Conf. Interval]

age                           0.1392668   0.0363307  3.83   0.000    .0679081  0.2106255
1.idicatorone                 0.4070426   0.2380546  1.71   0.088   -.0605304  0.8746155
1.indicatortwo                1.085003    0.1235029  8.79   0.000    .8424262  1.327581
1.indicatorthree              -0.4390229  0.1781991  -2.46  0.014   -.7890312 -0.0890145
1.fincrisis                   0.4526393   0.1376538  3.29   0.001    .1822677  0.7230108
1.bubble                      1.657902    0.1088199  15.24  0.000    1.444165  1.87164
indicatorone#indicatortwo
1 1                       0.0810565   0.2113363  0.38   0.701   -.3340379  0.496151
bubble#indicatorthree
1 1                       -0.8225963  0.2688304  -3.06  0.002   -1.350617 -0.2945752
fincrisis#indicatorthree
1 1                       -0.1068982  0.2141135  -0.5   0.618   -.5274474  0.313651
indicatorthree#c.age
1                             -0.0103719  0.0450553  -0.23  0.818   -.0988669  0.0781231
_cons                        0.0378515   0.1878819  0.2    0.840   -.3311753  0.4068783

• What is the exact command that resulted in that message? Commented Aug 21, 2013 at 13:41
• When I use the regress command, the F test does not show. I click on it, since its in blue, and it leads me to help file which shows that message. Commented Aug 21, 2013 at 13:43
• Can you paste the entire regression output in your question? Commented Aug 21, 2013 at 13:46
• Done, I'm not sure how to get it in a more presentable format, but that's basically it. Commented Aug 21, 2013 at 13:53
• Has anything been dropped from this regression due to perfect collinearity? 9+569 do not add up to 580, and you have 10 predictors rather than 9 used for the F-statistic. Commented Aug 21, 2013 at 17:28