You are being given the covariance matrix of a multiple regression
with where are numerical
explanatory variables and
are 4 binary dummy variables for 4 different categories
relative to the baseline category.
The categorical explanatory variable has 5 categories.
You are given that the estimated covariance matrix is:
or in R:
acov=matrix(c(0.33347, -0.00123, 0.000379, -0.00635, -0.007631, -0.000744, -0.036518, -0.054614, -0.049256, -0.043163, -0.00123, 1.1e-05, 3e-06, 0, 3.6e-05, 1e-06, 1.9e-05, 0.000121, 0.0002, 9.9e-05, 0.000379, 3e-06, 1.9e-05, -3.5e-05, 2.6e-05, -1.1e-05, -7.2e-05, 1.2e-05, 3.2e-05, -1.6e-05, -0.00635, 0, -3.5e-05, 0.000281, 7.7e-05, 1.9e-05, 0.000796, 0.000944, 0.000292, 0.000309, -0.007631, 3.6e-05, 2.6e-05, 7.7e-05, 0.00051, -4e-05, 0.001221, 0.002007, 0.001827, 0.001396, -0.000744, 1e-06, -1.1e-05, 1.9e-05, -4e-05, 1.6e-05, -3.2e-05, -6.9e-05, -2.1e-05, 6.8e-05, -0.036518, 1.9e-05, -7.2e-05, 0.000796, 0.001221, -3.2e-05, 0.022795, 0.016794, 0.014729, 0.014576, -0.054614, 0.000121, 1.2e-05, 0.000944, 0.002007, -6.9e-05, 0.016794, 0.030437, 0.017899, 0.016878, -0.049256, 0.0002, 3.2e-05, 0.000292, 0.001827, -2.1e-05, 0.014729, 0.017899, 0.025177, 0.017092, -0.043163, 9.9e-05, -1.6e-05, 0.000309, 0.001396, 6.8e-05, 0.014576, 0.016878, 0.017092, 0.029926),10,10)
Please use 3 decimal places for the answers below which are not
integer-valued.
Part a)
Based on the (estimated) covariance matrix,
what is the SE of ?
Note that corresponds to
the estimated distance of the hyperplanes for two different categories
where the binary dummy variables are indexed as and
.
Part b)
Based on the (estimated) covariance matrix,
what is the SE of ?
Note that corresponds to
a contrast to two categories (where the binary dummy variables are indexed
as ) and
with a third category (where the binary dummy variable is
indexed as ).
Hint:
You can earn partial credit on this problem.