Econometrics summary
Is the LPM model appropriate? Yes, because no predicted values <0 or >1 (LPM is
for dummy variables as an independent variable). Next to this you have to use robust
to correct for heteroskedasticy.
If the variable in the upper left corner in the table is a dummy, you use percentage
points.
Question about LPM: answer with use robust for heteroskedasticity. If it goes outside
the range of 0-1 it is not good.
When you have a level and a dummy you answer in hours or the given thing you do
NOT use percentages. ONLY with log you answer in percentages.
? > 0 it is a minimum
? < 0 it is a maximum
Dynamic model = with a Y variable
Static model = only x variable
Unit root:
TriangleY = a + /0Xt-1 + (a1t) +Et
H0: B1= 0 unit root
Ha: B1< 0 no unit root
Cointergration:
do the same for the unit root except different t value
H0: P = 0 no cointegration
Ha: P < 1 cointegration
Autocorrelation:
H0: p = 0 autocorrelation
Ha: p /= 0 no autocorrelation
No first order auto correlation if P > 0.05
A1. The regression model is linear, is correctly specified, and has an additive error
2
term. Explanation: no squared parameters 𝛽 or multiplication of the parameters 𝛽0 ×
𝛽1. So the beta itself cannot have a squared function or can be multiplied but the
term can.
This is correct: B2expr2
This is not correct B^2expr
Is the LPM model appropriate? Yes, because no predicted values <0 or >1 (LPM is
for dummy variables as an independent variable). Next to this you have to use robust
to correct for heteroskedasticy.
If the variable in the upper left corner in the table is a dummy, you use percentage
points.
Question about LPM: answer with use robust for heteroskedasticity. If it goes outside
the range of 0-1 it is not good.
When you have a level and a dummy you answer in hours or the given thing you do
NOT use percentages. ONLY with log you answer in percentages.
? > 0 it is a minimum
? < 0 it is a maximum
Dynamic model = with a Y variable
Static model = only x variable
Unit root:
TriangleY = a + /0Xt-1 + (a1t) +Et
H0: B1= 0 unit root
Ha: B1< 0 no unit root
Cointergration:
do the same for the unit root except different t value
H0: P = 0 no cointegration
Ha: P < 1 cointegration
Autocorrelation:
H0: p = 0 autocorrelation
Ha: p /= 0 no autocorrelation
No first order auto correlation if P > 0.05
A1. The regression model is linear, is correctly specified, and has an additive error
2
term. Explanation: no squared parameters 𝛽 or multiplication of the parameters 𝛽0 ×
𝛽1. So the beta itself cannot have a squared function or can be multiplied but the
term can.
This is correct: B2expr2
This is not correct B^2expr
