Dear Statalists
I'm in the final phase of my research and I want to look into the interaction effect of two variables.
I'm having trouble with multicollinearity because of the interaction between one of the two variables and the interaction variable.
I know mean centering variables is often used with this problem.
However, I must be doing something wrong because my variable doesn't change after substracting the mean at every observation as well as my vif values do not change.
I didn't know how to solve this in one step so I tried doing it with multiple steps.
1) I asked for the mean value of the interaction variable
2) I generated a new variable with this value in every cell/observation
3) I substracted the initial interaction variable by the mean of the variable
Am I doing something wrong here? Can someone please help me out.
My coding process is below.
Kind regards
Kletser Jok
I'm in the final phase of my research and I want to look into the interaction effect of two variables.
I'm having trouble with multicollinearity because of the interaction between one of the two variables and the interaction variable.
I know mean centering variables is often used with this problem.
However, I must be doing something wrong because my variable doesn't change after substracting the mean at every observation as well as my vif values do not change.
I didn't know how to solve this in one step so I tried doing it with multiple steps.
1) I asked for the mean value of the interaction variable
2) I generated a new variable with this value in every cell/observation
3) I substracted the initial interaction variable by the mean of the variable
Am I doing something wrong here? Can someone please help me out.
My coding process is below.
Kind regards
Kletser Jok
Code:
. gen Interaction= UAI* Altman_high
. reg CAR_win Interaction UAI Altman_high LnSize Leverage LnDealValue Domestic Horizontal Japan US
Source | SS df MS Number of obs = 176
-------------+---------------------------------- F(10, 165) = 1.80
Model | .038689375 10 .003868937 Prob > F = 0.0650
Residual | .355554565 165 .002154876 R-squared = 0.0981
-------------+---------------------------------- Adj R-squared = 0.0435
Total | .394243939 175 .002252823 Root MSE = .04642
------------------------------------------------------------------------------
CAR_win | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Interaction | -.0000212 .0000375 -0.56 0.574 -.0000953 .000053
UAI | .0004438 .000323 1.37 0.171 -.000194 .0010816
Altman_high | .0003951 .0023014 0.17 0.864 -.0041489 .004939
LnSize | -.0047945 .0023444 -2.05 0.042 -.0094233 -.0001657
Leverage | -.0230251 .0208805 -1.10 0.272 -.0642525 .0182023
LnDealValue | .0052026 .0024621 2.11 0.036 .0003412 .010064
Domestic | -.0058679 .0095064 -0.62 0.538 -.0246378 .0129019
Horizontal | -.0071467 .0072896 -0.98 0.328 -.0215395 .0072462
Japan | -.0237803 .0124236 -1.91 0.057 -.04831 .0007493
US | -.0265029 .0131432 -2.02 0.045 -.0524535 -.0005522
_cons | .0488307 .0304801 1.60 0.111 -.0113505 .1090119
------------------------------------------------------------------------------
. vif
Variable | VIF 1/VIF
-------------+----------------------
Interaction | 10.29 0.097223
Altman_high | 10.03 0.099713
UAI | 4.12 0.242708
LnSize | 3.70 0.270312
LnDealValue | 3.29 0.304259
Japan | 2.76 0.362260
Leverage | 1.33 0.753473
US | 1.30 0.771970
Domestic | 1.15 0.869232
Horizontal | 1.08 0.921773
-------------+----------------------
Mean VIF | 3.90
. mean (Interaction)
Mean estimation Number of obs = 176
--------------------------------------------------------------
| Mean Std. Err. [95% Conf. Interval]
-------------+------------------------------------------------
Interaction | 243.902 22.60059 199.2972 288.5068
--------------------------------------------------------------
. gen Meaninteraction= 243.902
. gen MC_Interaction=Interaction- Meaninteraction
. reg CAR_win MC_Interaction UAI Altman_high LnSize Leverage LnDealValue Domestic Horizontal Japan US
Source | SS df MS Number of obs = 176
-------------+---------------------------------- F(10, 165) = 1.80
Model | .038689375 10 .003868937 Prob > F = 0.0650
Residual | .355554565 165 .002154876 R-squared = 0.0981
-------------+---------------------------------- Adj R-squared = 0.0435
Total | .394243939 175 .002252823 Root MSE = .04642
--------------------------------------------------------------------------------
CAR_win | Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------------+----------------------------------------------------------------
MC_Interaction | -.0000212 .0000375 -0.56 0.574 -.0000953 .000053
UAI | .0004438 .000323 1.37 0.171 -.000194 .0010816
Altman_high | .0003951 .0023014 0.17 0.864 -.0041489 .004939
LnSize | -.0047945 .0023444 -2.05 0.042 -.0094233 -.0001657
Leverage | -.0230251 .0208805 -1.10 0.272 -.0642525 .0182023
LnDealValue | .0052026 .0024621 2.11 0.036 .0003412 .010064
Domestic | -.0058679 .0095064 -0.62 0.538 -.0246378 .0129019
Horizontal | -.0071467 .0072896 -0.98 0.328 -.0215395 .0072462
Japan | -.0237803 .0124236 -1.91 0.057 -.04831 .0007493
US | -.0265029 .0131432 -2.02 0.045 -.0524535 -.0005522
_cons | .043672 .0328617 1.33 0.186 -.0212115 .1085556
--------------------------------------------------------------------------------
. vif
Variable | VIF 1/VIF
-------------+----------------------
MC_Interac~n | 10.29 0.097223
Altman_high | 10.03 0.099713
UAI | 4.12 0.242708
LnSize | 3.70 0.270312
LnDealValue | 3.29 0.304259
Japan | 2.76 0.362260
Leverage | 1.33 0.753473
US | 1.30 0.771970
Domestic | 1.15 0.869232
Horizontal | 1.08 0.921773
-------------+----------------------
Mean VIF | 3.90
.
Comment