Linear models

True, False, Uncertain. Explain your answer. Credit entirely depends on the explanation.

For these propositions, assume that the observable variable > is determined by observable variables X and unobservable variable & via the linear relationship y = XB +E COV ( X, E ) = 0 1. If the variables 1 and *2 obey the relationship *1 =(*2), then & is not identified. 2. If 1 2 , then data can reveal nothing about 1 and Pz 3. If there errors in this linear model are not exchangeable, then the parameter vector / is not identified. 4. If any of the observables are measured with error, then the parameter vector / is not identified. 5. If one of the variables in the vector X is omitted from the model, then the parameters associated with the other X variables are not identified. 6. Suppose that COV (X,, 8 ) = 0 cov(X2.=) =0 cov(X3.=) = 0 etc. In other words, the first regressor is correlated with the model error. If there exists a variable Z such that COV ( Z, E) =0 and cov (z, x) * 0 then B is identified