Recall the 2000 presidential election data. Suppose that we consider a regression of pBuch, which stands for proportion voting for Buchanan, as a response on the following variables as predictors: Inco: Mean personal income in 1994; Bla: Percentage of blacks in 1996; Hisp: Percentage of Hispanics in 1996; 265: Percentage of the population aged 65 and over; Coll: Percentage of the population that graduated from college based on 1990 census. Based on these predictors excluding the observations corresponding to palm beach county, suppose that we can construct the following linear combination of the predictors: .LiBERTOUN PALMBEACH -0.0004 -0.0002 0.0002 0.0004 Figure 2: Problem 3 1 X = 0.42 log[Bla] + 0.23 log[Hisp] + 0.44 log[2 65] - 0.75 log[Coll] - 0.15Inco-1. The OLS fit and its some graphical diagnostics for a regression of log-scaled pBuch on X is given in Figure 2. Figure 2 supports that the OLS fit looks good for the regression. In Figure 2, the dashed lines represent the OLS fit, and the solid lines do LOWESS smoothing with fixing the tunning parameter at 0.75. Now the palm beach county observations are restored and pointed in the figure along with the two other counties of Calhoun and Liberty. And the 95% confidence intervals for their predictions are as follows: . Palm Beach County - Actual votes for Buchanan: 3407 Predicted votes: 1333.26; 95% CI [562.30, 3161.31] . Calhoun County - Actual votes for Buchanan: 90 Predicted votes: 20.25; 95% CI [8.56, 47.958] . Liberty County - Actual votes for Buchanan: 39 Predicted votes: 9.65; 95% CI [4.08, 22.848] Based on these results including Figure 2, what can you conclude for deciding whether or not palm beach county is an outlier or not