SPH 7460 Final Exam December 17, 2011 page 1 of 7 Six Problems - 7 pages Name: _________________________________________ ===================================================================================== 1. Let X = person's height and Y = person's serum creatinine level. Suppose you have a sample of N people and that X and Y are measured on each person. a) How would you construct a test for the hypothesis that corr(X, Y) = 0? That is, write a program that would do this without using a SAS procedure like PROC CORR or an analogous R procedure. You can assume that X and Y are approximately normally distributed. [5] b) Write a program using the bootstrap that will produce a 95% confidence interval for the true correlation. [Continue on next page if needed] [10] SPH 7460 Final Exam December 17, 2011 page 2 of 7 Name: _________________________________________ ===================================================================================== SPH 7460 Final Exam December 17, 2011 page 3 of 7 Name: _________________________________________ ===================================================================================== 2. A triangle T has vertices (0, 0), (3, 2), and (5, 1). a) Find a linear transformation F: R^2 --> R^2 such that F(T) is the triangle with vertices (0, 0), (1, 0), and (0, 1). [7] b) Use part a) to find the area of the original triangle T. [6] SPH 7460 Final Exam December 17, 2011 page 4 of 7 Name: _________________________________________ ===================================================================================== 3. An observation X in a dataset is called an OUTLIER if X > Q3 + 1.5 * (Q3 - Q1) or X < Q1 - 1.5 * (Q3 - Q1), where Q3 is the third quartile of dataset (i.e., 3/4 of the observations are less than Q3) and Q1 is the first quartile. Write a macro which lists all the outliers for a given variable in a dataset. The call to the macro should be of the form %outliers (indata, xvar, outdata), where indata is the input dataset, xvar is the variable of interest, and outdata is the output dataset of outlying values. [20] SPH 7460 Final Exam December 17, 2011 page 5 of 7 Name: _________________________________________ ===================================================================================== 4. Given the following data: Obs X --- ----- 1 13 2 4 3 1 4 11 5 1 6 12 7 3 8 5 9 10 10 2 You need to decide whether these data came from a Poisson distribution or some other distribution. How would you do this? You can use any software you want - SAS procedures, R, etc.. The important thing is to describe how you would approach the problem. [20] SPH 7460 Final Exam December 17, 2011 page 6 of 7 Name: _________________________________________ ===================================================================================== 5. A program produces maximum likelihood estimates theta_hat and phi_hat for two parameters, theta and phi. The program also computes the the standard error of theta_hat, the standard error of phi_hat, and the covariance of theta_hat and phi_hat. These are: std err (theta_hat) = .5 std err (phi_hat) = .3 covar (theta_hat, phi_hat) = .2 Find the estimated standard error of: g(theta_hat, phi_hat) = theta_hat * phi_hat + cos(theta_hat + phi_hat). [16] SPH 7460 Final Exam December 17, 2011 page 7 of 7 Name: _________________________________________ ===================================================================================== 6. Let f(x) = exp(x) + exp(-x). Write a program using Newton's method that will find a solution x to the equation f(x) = 3, [16]