Need professional help!!! Thanks in advance!! Please refer to Volume 1 (page 348) question 28. It’s from 2006 real Level II exam. My question is: If we are talking about “parameter estimate uncertainty” and “regression model uncertainty”, what test results should we look into? The solution on the book is not clear. For me, I think (1) t-test for individual coefficient is a good candidate for “parameter estimate uncertainty” , i.e. if it is significant, then it is certain. (2) F-test is a good candidate for “regression model uncertainty” , i.e. if it is significant, then the model is certain. Any thoughts??
SEE of the model and of parameter estimates, as long see s are present the model/estimates will be uncertain , also rsq. for the model
I agree with your choices, but the “if it’s significant then it’s certain” stuff is bogus. If it’s significant, then the possibilities are: a) There is some relationship among the variables giving you the significant test statistic b) An unusual event has happened when there is no relationship c) The assumptions of your test are wrong d) The sampling got messed up somehow (like you don’t have random sampling). or probably a bunch of other possibilities…
Ok. Thanks guys! But who can explain that why the answer is D for this particular question?
errata maybe…I see your point. lemme look
Actually, the answer is C. But I chose D. Any thoughts?
usrich, maybe you can help us out by clarifying what part of this question is causing you confusion. I’ll reprint the question and answer for the benefit of others. p.346 (excerpt from vignette): After responding to her intern’s questions, Chiesa concludes with the following statement: “Predictions from Exhibit 1 are subject to parameter estimate uncertainty, but not regression model uncertainty.” Question 28: Is Chiesa’s concluding statement correct regarding parameter estimate uncertainty / regression model uncertainty A. no/no B. no/yes C. yes/no D. yes/yes Appendix (A-27): C is correct. Predictions in a multiple regression model are subject to both parameter estimate uncertainty and regression model uncertainty (std. error of the estimate).
usrich Wrote: ------------------------------------------------------- > Actually, the answer is C. But I chose D. Any > thoughts? That’s a grammar issue: Chiesa concludes with the following statement: “Predictions from Exhibit 1 are subject to parameter estimate uncertainty, but not regression model uncertainty.” Chiesa thinks that parameter estimate uncertainty is an issue but that regression model uncertainty is NOT an issue. He is correct in thinking that parameter estimate uncertainty is an issue. He is NOT correct in thinking that regression model uncertainty is NOT an issue.
Hi guys, just to back to a post on the same vignette; in Question 26, Page 359 of Volume 1, Practice Problems for Reading 12…why is the answer B and not C? The solution says to use the p for 0.025, which gives a t-stat of 1.96 whereas the question asks for the 95% confidence interval…shouldn’t this mean take the p for 0.05? Thanks!