Hi all,

I have a dependent variable which varies between 0 and 1 so I decided to use fracreg. However, the pseudo r-square is too low comparing to the one provided by a regular OLS. I use fracreg to make sure that all the fitted values vary in the correct unit, 0 and 1. I'm leaving the output of regress and fracreg, respectively.

Code:
Linear regression                               Number of obs     =      2,553
                                                F(29, 247)        =      28.74
                                                Prob > F          =     0.0000
                                                R-squared         =     0.4298
                                                Root MSE          =     .17671

                                      (Std. Err. adjusted for 248 clusters in CompanyNo)
----------------------------------------------------------------------------------------
                       |               Robust
       MarketLeveraget |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
         Tangibilityt1 |    .419963   .0402887    10.42   0.000     .3406098    .4993161
       Profitabilityt1 |  -.3345266   .0690199    -4.85   0.000    -.4704692    -.198584
   MarkettoBookRatiot1 |  -.0516257   .0136449    -3.78   0.000    -.0785009   -.0247506
   OperatingLeveraget1 |  -.0481949    .012368    -3.90   0.000    -.0725552   -.0238346
           AssetRiskt1 |  -.1730909   .0498386    -3.47   0.001    -.2712537   -.0749281
DividendPayingStatust1 |  -.0499783   .0139422    -3.58   0.000     -.077439   -.0225175
     Sizelogt1Adjusted |   .0263341   .0134904     1.95   0.052    -.0002368    .0529049

Code:
Fractional logistic regression                  Number of obs     =      2,553
                                                Wald chi2(29)     =     770.97
                                                Prob > chi2       =     0.0000
Log pseudolikelihood = -1627.1735               Pseudo R2         =     0.0760

                                      (Std. Err. adjusted for 248 clusters in CompanyNo)
----------------------------------------------------------------------------------------
                       |               Robust
       MarketLeveraget |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
         Tangibilityt1 |   1.882211   .2193187     8.58   0.000     1.452354    2.312068
       Profitabilityt1 |  -1.873641    .362013    -5.18   0.000    -2.583174   -1.164109
   MarkettoBookRatiot1 |  -.4021407   .0702935    -5.72   0.000    -.5399134   -.2643679
   OperatingLeveraget1 |  -.3152975    .082395    -3.83   0.000    -.4767887   -.1538062
           AssetRiskt1 |  -.7806109   .2395676    -3.26   0.001    -1.250155   -.3110669
DividendPayingStatust1 |  -.1927583   .0641794    -3.00   0.003    -.3185476    -.066969
     Sizelogt1Adjusted |   .1237303   .0594861     2.08   0.038     .0071398    .2403208
Should I even report the pseudo R-square? Does it really make sense to make an inference on the goodness of fit of the model by considering pseudo R2?

Many thanks.