Hi,

I would like to ask for some clarification on regression interpretation.

The dependent variable is the number of child abuse in rate per 1,000 and the independent variables are the amount of subsidy ($/per child) and the percentage of White population (%).

I would like to conduct a regression model.

Code:
. reg ChildAbuse subsidy White

      Source |       SS           df       MS      Number of obs   =       153
-------------+----------------------------------   F(2, 150)       =      0.19
       Model |  9.78712136         2  4.89356068   Prob > F        =    0.8231
    Residual |  3765.58673       150  25.1039116   R-squared       =    0.0026
-------------+----------------------------------   Adj R-squared   =   -0.0107
       Total |  3775.37386       152  24.8379859   Root MSE        =    5.0104

------------------------------------------------------------------------------
  ChildAbuse |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     subsidy |   .0001577   .0002916     0.54   0.589    -.0004184    .0007338
       White |    .007741   .0388029     0.20   0.842    -.0689299    .0844119
       _cons |   9.235017   .7771175    11.88   0.000     7.699506    10.77053
------------------------------------------------------------------------------
Since the coefficient for subsidy is low, I converted into log format to control the effect.

then, I get:

Code:
. reg ChildAbuse lnsubsidy White

      Source |       SS           df       MS      Number of obs   =       153
-------------+----------------------------------   F(2, 150)       =      0.17
       Model |  8.40519016         2  4.20259508   Prob > F        =    0.8461
    Residual |  3766.96867       150  25.1131244   R-squared       =    0.0022
-------------+----------------------------------   Adj R-squared   =   -0.0111
       Total |  3775.37386       152  24.8379859   Root MSE        =    5.0113

------------------------------------------------------------------------------
  ChildAbuse |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   lnsubsidy |  -.3793123   .7784218    -0.49   0.627      -1.9174    1.158776
       White |   .0128794   .0381096     0.34   0.736    -.0624217    .0881804
       _cons |   12.32975   5.818937     2.12   0.036     .8320792    23.82742
------------------------------------------------------------------------------
Due to the units, I am a bit confused interpreting these results.

For the previous variable, would it be: "If I increase a dollar amount of subsidy, it decreases the number of child abuse (per 1,000) by -0.379, holding other variables constant."?

Thank you very much!