4. (Based on 9.16) The data set SOCWORK contains the salaries and years of experience for 50 social workers. (c) We learned one option for dealing with heteroscedastic residuals is to transform the dependent variable y. i. What transformation is suggested by the residual plot as well as context? ii. Fit a second-order model in experience for the transformation of salary you chose in part c(i). Confirm the residual plot appears homoscedastic. Is there evidence the degree two term is useful in this model? iii. Remove the degree two term from the model in c(ii) and re-fit. This will be Model 2. Note its R vs the model with a second-order term in c(ii) (These are comparable because you're predicting the same transformed y in both). EXP SALARY LNSALARY 1 7 26075 10.1687 28 79370 11.2819 2 3 4 23 65726 11.0932 18 41983 10.6450 un 19 62308 11.0398 6 15 41154 10.6251 7 24 53610 10.8895 13 33697 10.4252 8 9 10 2 22444 10.0188 8 32562 10.3909 11 20 43076 10.6707 12 21 56000 10.9331 13 18 58667 10.9796 14 7 22210 10.0083 15 2 20521 9.9292 16 18 49727 10.8143 17 11 33233 10.4113 18 21 43628 10.6835 19 4 16105 9.6869 20 24 65644 11.0920 21 20 63022 11.0512 22 20 47780 10.7744 23 15 38853 10.5675 24 25 66537 11.1055 25 25 67477 11.1195 26 28 64785 11.0788 27 26 61581 11.0281 28 27 70678 11.1659 29 20 51301 10.8455 30 18 39346 10.5801 31 1 24833 10.1199 32 26 65929 11.0963 33 20 41721 10.6388 34 26 82641 11.3223 35 28 99139 11.5043 36 23 52624 10.8709 37 17 50594 10.8316 38 25 53272 10.8832 39 26 65343 11.0874 40 19 46216 10.7411 41 16 54288 10.9021 42 L.. 20844 9.9448 43 12 32586 10.3916 44 23 71235 11.1737 45 20 36530 10.5059 46 19 52745 10.8732 47 27 67282 11.1166 48 25 80931 11.3014 49 12 32303 10.3829 50 11 38371 10.5551 4. (Based on 9.16) The data set SOCWORK contains the salaries and years of experience for 50 social workers. (c) We learned one option for dealing with heteroscedastic residuals is to transform the dependent variable y. i. What transformation is suggested by the residual plot as well as context? ii. Fit a second-order model in experience for the transformation of salary you chose in part c(i). Confirm the residual plot appears homoscedastic. Is there evidence the degree two term is useful in this model? iii. Remove the degree two term from the model in c(ii) and re-fit. This will be Model 2. Note its R vs the model with a second-order term in c(ii) (These are comparable because you're predicting the same transformed y in both). EXP SALARY LNSALARY 1 7 26075 10.1687 28 79370 11.2819 2 3 4 23 65726 11.0932 18 41983 10.6450 un 19 62308 11.0398 6 15 41154 10.6251 7 24 53610 10.8895 13 33697 10.4252 8 9 10 2 22444 10.0188 8 32562 10.3909 11 20 43076 10.6707 12 21 56000 10.9331 13 18 58667 10.9796 14 7 22210 10.0083 15 2 20521 9.9292 16 18 49727 10.8143 17 11 33233 10.4113 18 21 43628 10.6835 19 4 16105 9.6869 20 24 65644 11.0920 21 20 63022 11.0512 22 20 47780 10.7744 23 15 38853 10.5675 24 25 66537 11.1055 25 25 67477 11.1195 26 28 64785 11.0788 27 26 61581 11.0281 28 27 70678 11.1659 29 20 51301 10.8455 30 18 39346 10.5801 31 1 24833 10.1199 32 26 65929 11.0963 33 20 41721 10.6388 34 26 82641 11.3223 35 28 99139 11.5043 36 23 52624 10.8709 37 17 50594 10.8316 38 25 53272 10.8832 39 26 65343 11.0874 40 19 46216 10.7411 41 16 54288 10.9021 42 L.. 20844 9.9448 43 12 32586 10.3916 44 23 71235 11.1737 45 20 36530 10.5059 46 19 52745 10.8732 47 27 67282 11.1166 48 25 80931 11.3014 49 12 32303 10.3829 50 11 38371 10.5551