There are a few variables that would explain monthly income of households. Use the lookfor command to find the appropriate one.
lookfor income
267. farmrent int %9.0g crop rental income 268. liverent int %9.0g grazing rental income 269. rentinc float %9.0g rental income 281. agincome float %9.0g value of ag. income 283. otherinc float %9.0g household other income 285. totminc float %9.0g total monthy income
Now, try the regression as follows:
sort hhid
reg mxtfood totminc if hhid~=hhid[_n-1]
Source | SS df MS Number of obs = 1026 ---------+------------------------------ F( 1, 1024) = 65.37 Model | 10156513.2 1 10156513.2 Prob > F = 0.0000 Residual | 159108791 1024 155379.679 R-squared = 0.0600 ---------+------------------------------ Adj R-squared = 0.0591 Total | 169265304 1025 165136.882 Root MSE = 394.18 ------------------------------------------------------------------------------ mxtfood | Coef. Std. Err. t P>|t| [95% Conf. Interval] ---------+-------------------------------------------------------------------- totminc | .021563 .0026671 8.085 0.000 .0163295 .0267966 _cons | 493.7278 13.30798 37.100 0.000 467.6138 519.8419 ------------------------------------------------------------------------------
The equation made by these coefficients equals:
(predicted mxtfood)i= 493.7278 + .0211563(totminc)iFor every 1000 rand increase in total monthly household income, monthly household food expenditure increases by approximately 20 rand. From the regression table, t-values for constant and totminc both exceed 2, which implies these coefficients are significant. Also, R-squared value implies that variance of monthly food expenditure explained by total monthly income is 6%. Keep in mind, though, that we have not checked for outliers.