Question

. Necessary assumptions for regression two scatter plots and the corresponding regression lines in the following dilagrams. Identily which graph is more homoscedastic. Graph I Graph II Y SCORES Y SCORES 10 10 2 0 2&1 X SCORES 10 X SCORES You wish to use a least squares regression line to predict a new response value Y for a given explanatory value X. Which of the assumptions are necessary in order to expect some accuracy in the prediction? Check all that apply The relationship between X and Y must be nonlinear. The relationship between X and Y must be linear. The variability of Y must have a linear relationship with x. □ You can use the rngression I.ne to predict scorns only for thoso t n line to predict scores only for those that were originally in the sample used to compute the regression line

Answer 1

The scatter plots have points being represented on them. We must identify which graph (graph 1 or graph 2) is more homoscedastic.

For the points to be homoscedastic, they must be equally spread out above and below the regression line.

Homoscedasticity is basically the assumption that the variance of the error terms is constant. For this to be true, there must be no visible pattern in the errors.

In the graph I, the pattern is clearly visible. The errors in graph II have no visible patterns. Thus, the Graph II or the graph on the right is more homoscedastic.

If you have any doubts/ further points to add, please comment in the comments section, and I shall get back to you. Happy learning!