Hello,

I have run tobit regression with/without interaction terms.

Can I calculate interaction effect like this?

Why the sign of CR changed due to interaction terms?

what are the alternative ways of calculating interaction effect?

what is the difference between ##, and # for calculating interaction term?

Tobit regression
Div_Sales_W Coef. St.Err t-value p-value Sig.
ROA_W 0.100 0.002 45.05 0.000 ***
SGR_W -0.010 0.001 -16.09 0.000 ***
Size 0.002 0.000 14.48 0.000 ***
Equity 0.030 0.002 18.93 0.000 ***
Cash -0.030 0.002 -13.19 0.000 ***
RE -0.001 0.000 -15.99 0.000 ***
CR -0.123 0.008 -15.16 0.000 ***
JE_Time_log 0.169 0.005 33.65 0.000 ***
SR_2008 0.277 0.008 35.64 0.000 ***
c.SR_2008#c.JE_Tim_log -0.043 0.001 -34.04 0.000 ***
c.CR#c.JE_Time_log 0.021 0.001 16.05 0.000 ***
_cons -1.126 0.031 -36.51 0.000 ***
_cons 0.102 0.000 .b .b
Mean dependent var 0.030 SD dependent var 0.077
Pseudo r-squared -0.142 Number of obs 128350.000
Chi-square 11022.509 Prob > chi2 0.000
Akaike crit. (AIC) -88865.593 Bayesian crit. (BIC) -88738.680
*** p<0.01, ** p<0.05, * p<0.1

Tobit regression
Div_Sales_W Coef. St.Err t-value p-value Sig.
ROA_W 0.101 0.002 45.60 0.000 ***
SGR_W -0.011 0.001 -17.10 0.000 ***
Size 0.002 0.000 12.34 0.000 ***
Equity 0.030 0.002 18.62 0.000 ***
Cash -0.028 0.002 -12.41 0.000 ***
RE -0.001 0.000 -15.60 0.000 ***
CR 0.009 0.000 22.04 0.000 ***
JE_Time_log 0.018 0.001 23.59 0.000 ***
SR_2008 0.013 0.000 43.63 0.000 ***
_cons -0.197 0.005 -38.17 0.000 ***
_cons 0.102 0.000 .b .b
Mean dependent var 0.030 SD dependent var 0.077
Pseudo r-squared -0.127 Number of obs 128350.000
Chi-square 9861.251 Prob > chi2 0.000
Akaike crit. (AIC) -87708.335 Bayesian crit. (BIC) -87600.948
*** p<0.01, ** p<0.05, * p<0.1