Academic Integrity: tutoring, explanations, and feedback — we don’t complete graded work or submit on a student’s behalf.

Construct a 90% prediction interval for a single value of y for the following da

ID: 3177369 • Letter: C

Question

Construct a 90% prediction interval for a single value of y for the following data; use x = 100. Construct a 90% prediction interval for a single value of y for the following data; use x = 130. Compare the results. Which prediction interval is greater? Why? 90% prediction interval for a single value of y with x = 100: 90% prediction interval for a single value of y with x = 130: The width of this confidence interval of y for x_0 = is wider that the confidence interval of y for x_0 = because x_0 = is nearer to the value of x than is x_0 .

Explanation / Answer

Tabulate the calculations to determine the regression equation.

x   x-xbar   y   y-ybar   (x-xbar)(y-ybar)   (x-xbar)^2   ycap   (y-ycap)^2
141   59.6   25   -45.9   -2735.64 3552.16   17.590   54.9140
119   37.6   29   -41.9   -1575.44 1413.76   37.266   68.3334
103   21.6   44   -26.9   -581.04 466.56   51.577   57.4079
91 9.6 70 -0.9 -8.64 92.16   62.310   59.1423
63   -18.4 88   17.1   -314.64 338.56   87.353   0.4189
29   -52.4   112   41.1   -2153.64 2745.76   117.762   33.2053
24   -57.4   128   57.1   -3277.54 3294.76   122.234   33.2421

Using the above table, xbar=sigma x/n=570/7=81.4, ybar=496/7=70.9, sigma (x-xbar)(y-ybar)=-10647, sigma(x-xbar)^2=11904

Slope, beta1=sigma(x-xbar)(y-ybar)/sigma(x-xbar)^2=-10647/11904=-0.8944

Y intercept, beta0=ybar-beta1*xbar=70.9-(-0.8944)*81.4=143.7

Regression eqaution: yhat=143.7-0.8944x

Now, substitute x in the regression eqaution with given values of x and obtan ycap (see table). Subtract ycap from y and sqaure it to obtain SSE.

Therefore, standard error of residual, se=sqrt SSE/n-2=sqrt 306.66/7-2=sqrt 61.332=7.83

The 90% prediction interval is: ycapp+-talpha/2,df=n-2*se sqrt[1/n+(xp-xbar)^2/Sxx] [ycapp=143.7-0.8944*100=54.26, Sxx=sigma(x-xbar)^2]

=54.26+-2.015*7.83 sqrt[1/7+(100-81.4)^2/11904]

=47.7182, 60.8018

---

The 90% prediction interval is: ycapp+-talpha/2,df=n-2*se sqrt[1/n+(xp-xbar)^2/Sxx] [ycapp=143.7-0.8944*130=27.428, Sxx=sigma(x-xbar)^2]

=27.428+-2.015*7.83 sqrt[1/7+(130-81.4)^2/11904]

=18.2110, 36.6482