¦BASIC MPH WRITTEN EXAMINATION IN BIOSTATISTICSÈ ¦PART IIÈ ¦May 9, 1992: 9:30 am to 12:30 pmÈ ‚INSTRUCTIONSƒ: ™a)™This is an open book examination. ™b)™Answer any two questions during the three hour time period. ™c)™Put the answers to different questions on separate sets of papers. ™d)™Put your code letter, ‚notƒ your name, on each page. þì™e)™Return the examination with a signed statement of the honor pledge on a page separate from þìîyour answers. þìî™f)™You are required to answer only what is asked in the questions and not to tell all you know þìîabout the topics. þèîñõÞØÿÿÞØÿÿ e L š þìîö LQ1.™You have measured the glucose concentration (in mg/dl) in the anterior chamber of each eye of þî29 healthy dogs taken at random from a certain population, with results as follows: þì ™™Dog#™Right™Left™Dog#™%Right™,Left™4Dog#™;Right™BLeft ™™ 1™79™79™11™&77™-78™521™<71™C69 ™™ 2™81™82™12™&77™-77™522™<92™C93 ™™ 3™87™91™13™&84™-83™523™<91™C87 ™™ 4™85™86™14™&83™-82™524™;102™B101 ™™ 5™87™92™15™&74™-75™525™;116™B113 ™™ 6™73™74™16™&80™-80™526™<84™C80 ™™ 7™72™74™17™&78™-78™527™<78™C80 ™™ 8™70™66™18™%112™,110™528™<94™C95 ™™ 9™67™67™19™&89™-91™529™;110™B102 ™™ 10™69™69™20™&87™-91 þìYour research assistant has calculated the following sums, (uncorrected) sums of squares, and (uncorrected) sum of cross-products: ¦S(R)ÛV22439, S(L)ÛV22445, S(R*R)ÛV2209247, S(L*L)ÛV2210239, S(R*L)ÛV2209669È a)™™He tells you that a 95% confidence interval for the mean difference (RÛ,2L) extends from þîÛ,21.1 mg/dl to +0.7 mg/dl. Verify that he is correct. þî™b)™What you really wanted is a number X to substitute in the following sentence: With 95% þîconfidence, the difference in glucose concentration between the two eyes of the average dog in the population is less than X mg/dl." Your assistant tells you to use 1.1 for X, but you realize that 1.1 can't be right, since the mean difference in the sample is 1.7 mg/dl. For what slightly different sentence would 1.1 be correct? þî™c)™Explain to your research assistant how the X you wanted can be calculated. (Do not actually þîdo any calculations yourself.) þîQ2.™A group of four physicians hired a management consultant to see whether the long waiting time þîof their patients could be reduced. 200 patients were sampled randomly. It was found that their waiting time averaged 32 minutes with a standard deviation of 15 minutes. To determine the factors that waiting time depends on, a multiple regression was performed: ™™ ™™WAITÛV222Û+2.09 DRLATEÛ,2.24 PALATEÛ+22.61 SHORT R«2ÈÛV2.72 ™™™ standard error of the coefficient of DRLATEÛV2.01 ™™™ standard error of the coefficient of PALATEÛV2.05 ™™™ standard error of the coefficient of SHORTÛV2.82 ™™WAITÛV2waiting time in minutes ™™DRLATEÛV2the lateness of the doctors in arriving that morning (sum of their times in þîminutes) PALATEÛV2the lateness of the patient in arriving for the appointment in minutes SHORTÛV21 if the clinic was short staffed, and some of the appointments had to be rebooked, and it is 0 if fully staffed with all 4 physicians þî 10 pts.™™a)™ What conclusions can you draw? 15 pts.™™b)™ If you had been hired as a management consultant, how would you have proceeded with þî the analysis of the data? þî  Q3.™Your National Institutes of Health project monitor reviewed a paper that your authors wrote. Of þîthe 5,210 persons in the database, only 3,840 persons were analyzed for this paper because of various exclusion criteria. The monitor wants to know exactly why the 1,370 omitted persons were omitted. This problem asks you to write a SAS program to comply with his request. The criteria for inclusion follow: ™™¡i)™ The person must not be in the armed forces". ii)™™ The person must not have never worked". iii)™™ The person must not be a homemaker". iv)™™ The person must not be currently unemployed". v)™™ The person must have answered the work question as either working or retired. RESULTS A: The 1,370 excluded persons had a total of 2238 exclusion reasons. ™™59™ persons were in the armed forces" ™™™Á100™ persons had never worked" ™™™Á285™ persons were homemakers" ™™™Á501™ persons were currently unemployed" ™™™Â‚1293ƒ™ persons were neither working nor retired ™™™Â2238™ not necessarily mutually exclusive reasons for exclusion RESULTS B: As an aid to your understanding of this problem (you will not need these numbers or those above to write your program), here is the mutually exclusive categorization of exclusions: þì™™21™ persons never worked (only reason) ™™56™ persons were in the armed forces (only reason) ™™™Á496™ persons were neither working nor retired (only reason) ™™10™ persons were neither working nor retired AND they had never worked ™™™¡1™ person was neither working nor retired AND was in the armed forces ™™233™ persons were neither working nor retired and were homemakers þì™™52™ persons were neither working nor retired AND they were homemakers AND they had þì™™™ never worked ™™™Á482™ persons were neither working nor retired AND they were unemployed þì™™17™ persons were neither working nor retired AND they were unemployed AND they had þì™™™ never worked þì™™™¡2™ persons were neither working nor retired AND were unemployed AND were þì™™™Â›™ in the armed forces ™™™Â1370™ total persons excluded þìNote:™™™ The questionnaire, answers, and categorizations come from a real dataset. We didn't þî invent them, we only analyze them. þî  ™WHAT YOU WILL NEED TO WRITE YOUR PROGRAM ™1.™The SAS dataset (WORKQ) (5,210 observations) is stored online in a SAS datalibrary þîcatalogued as ™™™ DSNÛV2UFOOLU.BIG.STUDY.SASLIB. It is also stored on a PC-diskette in the subdirectory ™™™ A:\BIG\STUDY\SASLIB. ™™You may write your program either for the mainframe or for the PCÛ,2be sure to include the proper code for accessing the dataset, whichever platform you choose. þî ™2.™The variables used to establish the exclusion criteria are: ™™™ Q23:™9 =Û&2 never worked" ™™™ ™8 =Û&2 armed forces" ™™™ Q24:™3 =Û&2 homemaker" ™™™ ™6 =Û&2 unemployed" ™™™ ™if Q24 is not any of {1,2,4,5}, then the person is neither working nor retired" ™3.™Write a program which would have produced the numbers illustrated in RESULTS A:" and þîRESULTS B:" given above if it had used that dataset, but assume that the program must work with other samples from the same database. ATTENTION!!!!!! ™™The graders will award points based upon the following features of your program and answer: Û,2™ Correctly accessing the SAS dataset. Û,2™ Code to produce RESULTS A (the not necessarily mutually exclusive reasons for þî exclusion). þî ™™Û,2™ Code to produce RESULTS B (the mutually exclusive categorization of persons). ™™Û,2™ If necessary, a clear explanation of how the output of your program can be used to þî produce RESULTS A and RESULTS B. þî ™™Û,2™ Elegance, good programming practices (such as LABELs, FORMATs, LENGTHs, þî comments, TITLEs, etc., as appropriate), understandable code. þî  If you know exactly what you want your program to do but cannot devise the code, you may write in plain English (that a paid programmer might understand) what you would like to do and the logic of how to do it. If your explanation convinces the grader that your idea would work if implemented, you may get substantial partial credit. However, if you convince the grader that you don't understand what you are doing, you may be in trouble. Illegible answers will obtain very little credit.