Classic Classic Quiz Regression1 by Tomowsiak Tomowsiak 5 played - 12 yrs ago Regression1 QUIZ 11 QUESTIONS hard 1 An experiment is performed 3 times. Each time a model is fit. This means we generate how many slopes? How many y-intercepts? 1 intercept, 1 slope3 intercepts, 3 slopes 2 Using the question above, is it accurate to say that the slope and the y-intercept are both random variables? That is, they vary as we repeat our experiment and fit the model again and again? No. . they're the same every timeYes. . each experiment results in its own model. . they're similar though 3 When we fit a model, the line gives what exactly? Tells the average/mean response for a change of 1 unit along regressorTells us something elseTells us the exact value of each individual measurement from original data 4 The picture below shows what happens when we fit a model. Is this true or false? True. . for each x, we can generate many y's, which are normally distributedFalse. . this picture doesn't show what happens 5 It looks as though the two variables are col-linear. Which means what exactly? As one varies somehow, the other also varies. . so the two are bound. . so the response shows thisNothing. . they don't vary together. . the response reflects this independence 6 The hypothesis stated below are correct for a modeling using a multiple regressors. FalseTrue 7 The table shown below is correct. That is, it shows the correct entries for multiple regression. n-p=n-(k+1)=n-k-1 k is number of predictors TrueFalse 8 Prediction intervals are wider then confidence intervals because they depend on the current estimate of the mean and the spread of the data. That is, we fit a model and reach in a grab a new datapoint TrueFalse 9 If I conduct a regression analysis, and compute t stats. for each regression coefficient, then the square of these is what? It's nothing.It's the F value found from a partial regression test.It's something else. . 10 It's possible to conduct partial regression tests on the contribution made by several predictors, given that several others are already in the model. This is false.This is true. 11 I fit a model first with 1 regressor and then 2. R^2 adjusted increases. This means what? Nothing. . both are equally goodThis means the second regressor has reduced mean sqaure for error,It means something else A mistake in this Quiz ? Contact the author Commentaires