na multiple linear regression analysis, (y¡, X1i ,X2i), i = 1, ... ,n , are statistically ndependent and satisfy the model (M1) given by: y = Bo + B1x1 + B2X2 + E , here the response variable y is continuous, the regressor vector X = (x1,x,)' has mean ector (ur1, Hr2)' and positive definite variance-covariance matrix Ex, the random errors conditional on X; are statistically independent and normally distributed with mean zero nd variance o which does not depend on X . ) Consider the case that the value of o? is known; that is, it is given. 1) Construct a statistical test for testing Ho: B1 = B, = 0, using the known o², and the a-level rejection region of the test. [5 points] 2) Construct a statistical test for testing Ho: B, = B2, using the known o², and the a- level rejection region of the test. [5 points] 3) Are there any differences in testing between a1) and a2) above? why or why not? [5 points] ) Consider the case that the value of o? is unknown; that is, it needs to be estimated. 1) Construct a statistical test for testing Ho: B, = B2 = 0 and the a-level rejection region of the test. [5 points] 2) Construct a statistical test for testing each individual regressor; that is, for the i-th regressor (i = 1, 2), Hoi: ßi = 0 and the a-level rejection region of the test. [5 points] 3) In b2) above, derive the probability of falsely rejecting at least one Hoi , i = 1,2. In b1) above, derive the probability of falsely rejecting at least one Hoi , i = 1,2. [7 points] If Bo may or may not be zero in the model M1, construct analysis-of-variance (or sum-of- squares) table for this multiple linear regression analysis. [3 points]
In a multiple linear
? = ?0 + ?1?1 + ?2?2 + ? ,
where the response variable ? is continuous, the regressor
a) Consider the case that the value of ?2 is known; that is, it is given.
1) Construct a statistical test for testing ?0: ?1 = ?2 = 0, using the known ?2, and the ?-level rejection region of the test.
?0: ?1 = ?2 = 0
??: ?1 ≠ ?2 ≠ 0
2) Construct a statistical test for testing ?0: ?1 = ?2, using the known ?2, and the ?-level rejection region of the test.
3) Are there any differences in testing between a1) and a2) above? why or why not?
b) Consider the case that the value of ?2 is unknown; that is, it needs to be estimated.
1) Construct a statistical test for testing ?0: ?1 = ?2 = 0 and the ?-level rejection region of the test.
2) Construct a statistical test for testing each individual regressor; that is, for the i-th regressor (i = 1, 2), ?0i : ?i = 0 and the ?-level rejection region of the test.
3) In b2) above, derive the probability of falsely rejecting at least one ?0i, ? = 1, 2. In (b1) above, derive the probability of falsely rejecting at least one ?0i, ? = 1, 2.
c) If ?0 may or may not be zero in the model M1, construct analysis-of-variance (or sum-of-squares)
table for this multiple linear regression analysis.
Trending now
This is a popular solution!
Step by step
Solved in 2 steps with 1 images