illlllhldwl llli [I H ”H ’ ‘40) (/3010) AN- mmmmse as me munm OF Q-ussnmmmm mmwszs FENAENENG Tc; THE mommw as: Amzzm 71132.55 fat the Dawn 0% M. A. MBCHEGAN STAT’E UNE‘V’ERSS’E‘Y Em L. highway- W5 é AN INVESTIGATION OF THE VALIDITY OF QUESTIONNAIRE RESPONSES PERTAINING TO THE MORTALITY OF ATHLETES by EARL L. MAHONEY A THESIS Submitted to the College of Education of Michigan State University of Agriculture and Applied Science)in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Health, Physical Education, and Recreation 1956 ACKNOWLEDGMENT I wish to express my sincere gratitude to Dr. Henry Montoye for his help and guidance in preparation of this thesis. I am also indebted to Dr. Wayne VanHuss for his interest and suggestions; Mr. Robert Mooney of the Michigan State Public Health Department for his valuable assistance, and to my wife, Vera, for her patience and support through» out the preparation of this study. E.L.M. AN INVESTIGATION OF THE VALIDITY OF QUESTIONNAIRE RESPONSES PERTAINING TO THE MORTALITY OF ATHLETES by EARL L. MAHONEY AN ABSTRACT OF A THESIS Submitted to the college of education of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Health, Physical Education, and Recreation ,,H < 1956 > aszié (t) ~w~r Approved ,;[;,LL wccfwnflntg' w d EARL L. MAHONEY ABSTRACT Purpose g: the Study To investigate the validity of selected factual ques- tionnaire data from the Michigan State University Study of Longevity and Morbidity of Former College Athletes. Also, as a subordinate purpose, the interpretation and coding of the questionnaire causes of death by the investigators in the Michigan State University Study was studied. _Mgthodology Death certificates were used as the criteria to deter- mine the validity of the questionnaire data. The death cer- tifiCates were located for 104 of the 123 reported deaths in the Michigan State Study. The questionnaire data pertaining to date of birth, age of death, and cause of death were com- pared to the corresponding data on the death certificates. After the data were obtained, they were analyzed by conventional statistical techniques. Results The dates of birth as reported on 104 investigated questionnaires appear to be accurate. The correlation coef- ficient for the death certificates date of birth and the questionnaire date of birth was 0.99. The critical ratio was 0.75, for the difference in means. 2 EARL L. MAHONEY ABSTRACT The ages of death as reported on the 104 investigated questionnaires also appear to be accurate. The correlation for these data was 0.98, and the critical ratio was 0.13 for the difference in the means. A total of nineteen errors were observed when the questionnaire causes of death were compared to the death certificate causes of death. The errors constituted 18.3 per cent of the total 104 cases investigated. The causes of death appear to be valid data when classified in general categories. The nineteen errors were classified as to their magnitude. Two of the errors were classified as major error, thirteen as minor, and four as insufficient information. TABLE OF CONTENTS CHAPTER PAGE I. INTRODUCTION 0 O O O O O O O O O O O 1 Statement of the problem. . . . . . . . 2 Importance of the study . . . . . . . . 2 Limitations of the study. . . . . . . . 4 II. REVIEW or LITERATURE. . . . . . . . . . 6 Summary. . . . . . . . . . . . . . 11 III. METHODOLOGY . . . . . . . . . . . . . 13 Selection of data to be investigated. . . . 13 Investigative procedures. . . . . . . . 14 Classification of Causes of death. . . . . 16 Methods of analyzing the data . . . . . . 18 IV. RESULTS . . . . . . . . . . . . . . 21 Validation of the dates of death . . . . . 21 Validity of ages of death . . . . . . . 22 Validity of the causes of death . . . . . 22 Coding errors . . . . . . . . . . . 28 V. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS. . . 30 Summary. . . . . . . . . . . . . . 30 ConcluSionS . . . . . . . . . . . . 30 Recommendations. . . . . . . . . . . 31 BIBLIOGRAPHY. . . . . . . . . . . . . . . 33 APPENDIXES . . . . . . . . . . . . . . . 37 LIST OF TABLES TABLE PAGE I. A Comparison of Causes of Death as Reported on the Questionnaires and as Recorded on the Death Certificates . . . . . . . . . . 24 II. Comparison of All Causes of Death Before and After Correction of Coding Errors . . . . . 27 III. Effect of Coding Errors When Causes of Death are Coarsely Grouped. . . . . . . . . . 28 LIST OF FIGURES FIGURE PAGE 1. A Comparison of Causes of Death as Reported on the Questionnaires and as Recorded on the Death Certificates in Semi-Coarse Categories. . 25 2. A Comparison of Causes of Death as Reported on the Questionnaires anddas Recorded on the Death Certificates in Coarse Categories . . . 26 CHAPTER I INTRODUCTION The use of the questionnaire as an adequate investi- gative technique is in general acceptance by the social scientist. The assumptions basic to this technique must be taken for granted, for there are few objective data substan- tiating its validity. Data collected through the use of the questionnaire are subject to criticism, which is mainly centered about the reliability and the validity of the in- 1’2'3 Kelly,” however, calls attention to the fact strument. that many times the only instrument available for the col- lection of data, pertinent to a problem, is the questionnaire. Basically there were two types of data for which the questionnaire is employed: (1) Factual data, such as age, 1F. P. Whitney, “The Questionnaire Craze," Educational Review, 68:139-lh0, 1924. 2L. V. Koos, The ggestionnaire in Education (New York: The Macmillan Company, 1929), p. 2-3. 3Frederick L. Whitney, The Elements 9; Research (New York: Prentice-Hall, Inc.. 19575. p. 135. “T. L. Kelly, Scientific Method: Its Function ig Research and ig_Education (New York: The Macmillan Company, 1932;. p. 122. 2 sex, weight, and height; (2) Subjective data, such as opin- ions, attitudes, beliefs, and Values. The former are subject to verification, while the latter may or may not be subject to verification.5 In the Michigan State University Study of 6 the Longevity and Morbidity of Former College Athletes, more important data pertinent to final generalities involves factual data. gfiatement g£_the Problem It was the purpose of this study to determine the validity of selected factual questionnaire data of the Michigan State University Study of Longevity and Morbidityof Athletes and Non-Athletes. The selected data to be invest- igated for validity are cause of death, age of death, and date of birth. The criteria used to measure the degree of validity were Public Health Department records, namely, death certificates. Importance g; the Study The longevity and morbidity of former college athletes has long been a problem of extreme interest. As a result of this interest many investigators have conducted studies in 5George A. Lundberg, Social Research (New York: Longmans, Green and Co., 1942), p. 183. 6The Pilot Study of a National Study of Longevity and Morbidity of Athletes in Colleges and Universities, sponsored by Phi Epsilon Kappa fraternity, [a national, professional physical education honorary fraternity, inaugurated in 1951. this area.7 Some of the methods employed in the Michigan State University Study of Longevity and Morbidity of College Athletes were the results of an analysis of the methods other investigators have used. Some of the faults other investigators have made in this area of research.are poor sampling procedures, no control group, and no method to test the validity of the data. This study takes into consideration only one aSpect of the foregoing criticism, namely, validity. Although this study is concerned only with the validity of cause of death, age of death, and date of birth, the precise eValuation of these data, indirectly, pertains to the validity of thestudy as a whole. If the stigma of "unvalidated evidence,”yvhich usually accompanies most questionnaire studies, is removed the data becomes more meaningful. Another aspect of this study which may be taken into account is the development of an investigative technique, which may be used to validate certain factual data. Although the use of public records involves a great deal of tedious work, it is the belief of this investigator that a satis- factory systematic approach may be formulated to eXpedite this technique of investigation. 7Thomas K. Cureton, Physical Fitness Appraisal and Guidance (St. Louis: C. V. Mosby and Company, 1947), 309- 311. Limitations Q£_the Study In the classification of causes of death on the death certificates there might have been discrepencies due to the following reasons: (1) Physicians who practiced medicine thirty or forty years ago did not use the same methods of diag- nosis, nor the same nomenclature, as the physi- cians of today.8 (2) Many places in the United States have a county coroner who is not a professional medical man, and his certification as to cans of death is subject to question. (3) Medical entries on the death certificate are many times made to concurr with the legal sepects in regard to cause of death and are not concerned with the specific pathology involved.9 8Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death (Sixth revision; Geneva, Switzerland: World Healthfl Organization, 19u9), Vol. 1, p. xiii. 9Office of Vital Statistics, Physicians Handbook 2g Death and Birth Registration (Washington, D.C.: U. S. Government Printing Office, l9b9), p. A. (h) Difficulty in determining the most important cause of death, where two or more entries were made on the death certificate.10 Although the use of death certificates as validating criteria may be subject to the former criticisms, Dublin states: Registration of deaths is relatively accurate because of the legal importance of the event of death and the stress accordingly laid on the enforcement of certi- fication requirements. There was much variability among the death certificates in regard to basic design and methods of recording data due to differences in the various state health department regu- lations and chronological eras. Due to the lack of information, i.e., service number, rank, service unit, place and date of death; war deaths could not be validated. If these War deaths could have been investigated, it is assumed that the percentage of correct Questionnaire response would have been higher. 10Alan Rook, "An Investigation into the Longevity of Cambridge Sportsmen,” British Medical Journal, #865; 773-777, April, 1954. 11Louis I. Dublin, 31; a1, Length 2; Life (New York: Ronald. Press 00., 1914’9), p0 36-60 CHAPTER II REVIEW OF LITERATURE Much research has been done in regard to the longevity, morbidity, and mortality of athletes. A survey of the methods the various investigators used to gather their data is given. Morgan,1 in 1873, investigated the longevity of two hundred and ninetybfour former oarsmen from Oxford and Cambridge Universities. The data used in this study were records from the universities which are kept on all former students. In 1904 Meylan2 studied the effects of exercise and competition on the longevity and morbidity of former Harvard oarsmen. He attempted to see every former oarsman personally and examine him, when this was not possible he sent a card and asked the subject to have his family physician perform the examination. Thirty-two of his one hundred and fifty—two subjects had died. He stated the causes of death for twenty- two cases were definitely ascertained. The technique used to 1John E. Morgan, “Critical Inquiry into the After— Health of the Men Who Rowed in the Oxford and Cambridge Boat Race from the Year 1829-1859." University Oars, p. 330, cited by Peter V. Karpovich, "Longevity and Athletics," Research _Quarterly, Vol. 12, (May, 19U1), pp. 451-455. 2George L. Meylan, "Harvard University Oarsmen," Harvard figaduates Magazine, 9:355-356, March, 1904. 7 ascertain the causes of death was not mentioned in the study. He was unable to learn the exact cause of death in the other ten cases. 3 made a study of the longevity of former Yale Anderson athletes in 1916. With the aid of a statistician he secured his data from the records of the secretary of the university; i.e., files of the college publications, and obituary reports. Among the eight hundred and eight former athletes for which data were collected, there were only fifty-eight deaths. No attempt was made to validate the causes of death listed in the records, although they were reported as factual data. Greenway and Hiscocku made an analysis of the longe- vity and morbidity of former athletes and other graduates of Yale University in 1929. Again the class and obituary records, kept at the secretary's office, were used to gather the mat- erial for the investigation. In 1927, Hill5 studied the longevity of former cricket players in England. The source of data for his study was 3William G. Anderson, “Further Studies in the Long- evigy of Yale Athletes," Mind and Body, 23:374-375. December, 191 . James C. Greenway and Ira V. Hiscock, "Mortality Among Yale Men," Yale Alumni Weekly, 35:1086, June, 1926. 5A. Bradford Hill, "Cricket and Its Relation to the Duration of Life, "Lancet, 2:99h, 1927. "Wisden's Cricketers Almanac” which devoted a section to biographical details of former cricket players. Bickert6 carried out a study in Germany concerning the effects of competitive Sports on the longevity and cause of death of former champion athletes. Names of all his deceased subjects were collected from yearbooks and period- icals. Questionnaires were sent out to the deceased persons' athletic clubs. If the membership in an athletic club could not be determined, questionnaires were sent to the registrars' office of the communities in order to find the address of the person who had given the announcement of death. This person was then sent a questionnaire to answer questions concerning the deceased. The cause of death as reported in the study were not investigated for validity. Reed and Love7 made a study of the longevity of army officers in relations to physical fitness in 1931. The source of their data were the official physical and medical records of 5,021 individuals. The use of this type of data would lend a very high degree of validity to a study of this nature. ‘ 6 F. W. Bickert, "Einfluss, des Wettkampfmassig Betriebenen Sports Auf die Lebensdauer und Todesursache,” Deutsche Medizinische Wochenschrift. 55:23-24, 1929. 7Lowell J. Reed and Albert G. Love, “Longevity of Army Officers in Relation to Physical Fitness,” The Military Surgeion, 69:380, October, 1931. 9 Between the years 1938 and 19A0, Knoll8 and Llewellyn9 made independent investigations of the longevity of the former Oxford and Cambridge oarsmen. The centennial history of the rowing teams was used to collect the data for their studies. It is stated in the studies that this history con- tained exact data on the private lives and the ages of the participants. 1 In 1928,10 and again in 1932,1 Dublin made two in- vestigations related to the longevity of college athletes, from ten eastern universities. He used college records to secure his data for the first study. In the second study in which he used college graduates other than athletes as con- trols, actuary statistics from the Metropolitan Life Insur- ance Company were used to make comparisons. ,Presumably the data for both the athletes and controls were obtained from college records. This, however, is not clear in the report. 8W. Knoll, ”Welches Lebensalter Erreichen die Ruderer von 'Oxford-Cambridge'7' Medizinische Klinik, 3#:u64~#66, 1938. 9Percival H. Hartley and Geoffrey F. Llewellyn, "A Study of Those Who Rowed in the Oxford and Cambridge Boat Race from 1829 to 1928,” British Medical Journal, 1:658, April. 1939. 10Louis I. Dublin, "Longevity of College Athletes," Harpers Mgnthly Magazine, 157:230-231, 1928. 11Louis I. Dublin, "College Honor Men Long-Lived," Statistical Bulletin g; the Metropolitan Life Insurance Qp., 13:3‘69 19320 10 Wakefield,12 in 1944, completed a study of 2,900 former athletes who had played in the finals of the Indiana state high school basketball tournaments. His study was mostly concerned with the causes of mortality and age of death. The data for determining causes of death and ages of death were obtained from official sources such as State Board of Health Records, County Board of Health Records, Cemetary Records, and Records of Morticians. A study of the longevity of 400 former Czechoslovakian athletes was performed by Schmid in 1952.13 This study was mainly concerned with comparing the age of death of athletes with the normal life Span of the general population. The data were obtained through questionnaires sent out to relatives and friends of the deceased athletes. HOBpital archives and physicians records were also used as a source of data. In 1954, Rooklu studied the longevity of Cambridge oarsmen. He first used the "Alumni Cantabrigienses," a biographical list of former Cambridge students, to obtain such data as date of birth, economic, social, and when 12Mark C. Wakefield, “A Study of Mortality Among the Men Who Have Played in the Indiana High School State Final Basketball Tournaments," Research_guarter1y, 15:3-5. 1944. 13L. Schmid, "How Long the Sportsmen Live,"_§port and Health (Oslo, Norway: Royal Norwegian Ministry of Education, l“(Alan Rook, "An Investigation into the Longevity of Cambridge Sportsmen," British Medical Journal, 4865:777, April, 1954. 11 causes of death for one-half of the total group. No com- parisons were made of death certificate causes of death.with questionnaire causes of death. Rook also made the following statement in reference to the use of death certificates for the verification of causes of death: The use of death certificates as the basis for a statistical analysis of this nature is not entirely satisfactory and numerous difficulties are encountered. Diagnoses are often vague; fashions in diagnosis even on the death certificates change with the years, possibly capriciously or possibly as a result of more accurate methods and greater knowledge of pathological causes. Sometimes when two or more possible Causes of death figure on a certificatg it is difficult to choose the most important one.1 Summary In reviewing the studies concerned with the longevity of athletes there seems to be a definite change in trends regarding the methods used in collecting and analyzing the data. In the early studies the trend was to use college files as the primary source of data. In the later studies other methods of investigation have been used, Such as ques- tionnaires, actuary tables, and death certificates. In these later studies more emphasis was placed on the accuracy of the data. Also, the importance of ascertaining the 15Ib1d.. p. 777. 12 validity of the data cannot be over emphasized. Koos has stated: "It would be highly beneficial to the use of the questionnaire as an instrument of research if all those who use it would assume some responsibility for its validation."16 16L. V. Koos, The gpestionnaire ip_Education (New York: The Macmillan Company, 1929), p. 167. CHAPTER III METHODOLOGY The data which were investigated for validity were selected from the returned questionnaires of the Michigan State University Study of Longevity and Morbidity ofCollege Athletes.1 The questionnaires were from athletes and non- athletes, who had formerly attended Michigan State Univer- sity.2 The names and addresses of these former students were obtained from the files of the Athletic Director, Student Directories, The Registrars' Records, and from the Alumni Office. Questionnaires were sent out to 1,130 former ath- letes and 1,130 non-athletes.3 of the 2,260 questionnaires sent out, a total of 1,212 were returned with sufficient data 1H. J. Montoye, W. VanHuss, H. Olson, and A. Hudec, “A Study of the Longevity and Morbidity of College Athletes," unpublished study, Michigan State University, East Lansing, Michigan. Presented at the Third Annual Meeting of the "American College of Sports Medicine," March, 1956. 2As defined in Michigan State University study-~An athlete is one who has earned a major Sports letter and a control is a former student who attended M.S.U. during the time of attendance of the athlete, but who did not earn a major letter in any sport. 3See sample of questionnaire, Appendix A. 14 to be included in the study. The total percentage of returns, excluding those which were returned unopened due to incorrect addresses, etc., was 65.22 per cent. Of the 1,212 returned questionnaires 629 were from the athlete group and 583 from the control group. All the questionnaires which indicated the subjects to be deceased were separated from the questionnaires which denoted the subjects still living. The number of question- naires for the deceased group, which were filled out by relatives and friends of the deceased, numbered 123 or 9.85 per cent of the total returned questionnaires. There were sixty-seven athletes and fifty—six controls in this group. ‘Ipvestigative Procedures The dates of birth, ages of death, and causes of death were abstracted from the 123 questionnaires and tabulated.5 Four by six information cards were designed and printed, to be used in abstracting information from the death certifi- cateS.6 On the cards, blanks were left for information which “For further information of methods used in M.S.U. study refer to theses of: A. Hudec, "A Study of the Longe- vity and Morbidity of Football and Basketball Athletes at Michigan State University,” (unpublished Master's thesis, Michigan State University, East Lansing, Mich., 1956) passim. H. Olson, "A Study of Longevity and Morbidity of Track Ath- letes at Michigan State University, (unpublished Master's thesis, Michigan State University, East Lansing, Mich., 1956) p_a__ssim. 5See Appendix B for tabulation sheet. See Appendix C for sample of card. 15 appeared on both the questionnaire and the death certificate. Other entries, for information not appearing on the question- naire, such as race, occupation, and kind of business or industry, were also made. An abbreviated list of fifty causes of death for the tabulation of mortality was also entered.7 Five of the fifty causes of death on the original abbreviated list were eliminated as not being applicable to this study. The causes of death eliminated were those resulting from childbirth, birth injuries, and diseases peculiar to early infancy. Information regarding the subject's former state of residence, was obtained from the questionnaires. Where nec- essary, personal letters were sent to the individuals who had filled out the questionnaire to obtain more definite 8 information in regard to date and place of death. If the individual had been killed in the armed services, the branch of service and service number were requested. Personal letters and information cards were mailed to the appropriate Directors of the State Health Departments.9 The letter re- quested the information cards be completed in accordancewith the death certificate data, or that a non-certified cOpy of 1See Appendix D for abbreviated list of fifty causes of death cited from Manual 9£_the International Statistical Classification 2: Diseases, Injuries, and Causes g: Death (Sixth revision; Geneva, Switzerland: World Health Organiz- ation, 1949), Vol. 1, p. xiii. 8See Appendix E, letter. 9See Appendix F, letter. 16 the death certificate be returned in lieu of filling out the card. This information was provided free by all the states participating in the study except Mississippi. The Michigan State Department of Health made available to this investi- gator the death certificates of the subjects who had died in this state. Of the total number of subjects sixty-three were checked by the use of the Michigan State Health Department records. When the cards or death certificates were returned they were checked against the original questionnaire data. When there was disagreement, the death certificate data were tabulated in additional columns so that comparisons could be readily made. The most difficult part of the study was en- countered in the classification of causes of death. Classification _o___f: Causes <_)__f_‘_ Death The purpose of a statistical classification of causes of death is primarily to furnish quantitative data which will answer questions about groups of cases.10 The authority used for the classification of causes of death in this study and the Michigan State University Study was the Manual 22.22322? national Statistical Classificatippgp£_Diseases, Injuries, and Causes g_f_Death.11 This manual has three special lists 10Manual 2: the International Statistical Classifi- cation p£_Diseases, Injuries, and Causes 9: Death, op. cit.p.xi. 11Ibid., assim. 17 12 for tabulating morbidity and mortality. List B was used in the Michigan State University Study and the present study for classification of causes of death. In regard to methods used in classifying causes of death Farr states: Classification is a method of generalization. Several classifications may, therefore, be used with advantage; and the physician, the pathologist, or the jurist, each from his own point of view may legitimately classify the diseases and causes of death in the way that he thinks best adapted to facigitate his inquir- ies, and to yield general results.1 Early records of death usually contained a single cause, and only a few simple rules were in effect to secure uniform selection of the cause of death.14 Many of the death certi- ficates dealt with in this study contained multiple causes of death, and the problem of selecting one primary cause of death became very important. The authority used in this study stated that the underlying cause of death should be used in selecting the cause of death for classification purposes.15 This authority also stated: 13 Ibid., p. xii. 14 Ibid., p. xxxiv. 1 5Ibid., p. 345. 18 A cause of death is the morbid condition or disease process, abnormality, injury or poisoning leading directly, or indirectly, to death. Symptoms or modes of dying such as heart failure, asthenia, etc., are not considefgd to be causes of death for statistical purposes. The two principles which were followed in determining the primary cause of death were; (1) to use the disease or injury which initiated the train of morbid events leading directly, or indirectly to the cause of death; (2) or the circumstances of the accident or violence which produced the fatal accident.17 Methods 9: Analyzing_the Data When all the data had been collected and tabulated, the Pearson Product Moment method to determine the corre- lation coefficient of the questionnaire responses and the death certificate data, for the dates of birth and ages of death were used. The means were analyzed and the standard deviation and the critical ratio were computed for both the dates of birth and the ages of death data. The inconsistances which appeared between the ques- tionnaires.and.the death certificates regarding causes of death were recorded as errors. These errors were analyzed 19 in terms of percentages, magnitude, and the effect theerrors had on the total distribution of the causes of death. Three classifications were formulated for the grouping of errors. 1. Major Error—~Questionnaire response was not path- ologically related to the death certificate cause of death. 2. Minor Error-—Questionnaire reSponse was patholog- ically related to the death certificate cause, but would not be classified in the same specific category. 3. Insufficient Information-~Questionnaire reSponse did not give enough information for a separate classification of the cause of death. Tables and figures were made to show the distribution of the causes of death before and after correction of errors. Two separate categories of causes of death were formed; semi- coarse and coarse. The purpose of a semi-course and a coarse category of the data was to illustrate the effect the errors made when the Causes of death are grouped this way. These two categories are the most common used in reporting mortality data.19’20 18Tables and figures are presented in Chapter 4. 19Rook, 22. 931.. p. 777. 20L. I. Dublin, A. Lofka, M. Spiegelman, Length 2; Life (New York: Ronald Press, 1949), p. 83. 20 When the errors encountered in the questionnaire data were corrected on the original tabulation sheets, certain irregularities were found in the initial interpretations of the causes of death as reported on the questionnaires. The number and kind of erroneous interpretations were recorded. These coding errors were corrected in the originalciata. A table was made to compare the data before and after correction of the errorS. Another table was made which compared the data before the correction of coding errors, after the cor- rection of coding errors, with the causes of death coarsely grouped. CHAPTER IV RESULTS Validation g£_the Dates 9; Death The questionnaire dates of birth were compared with the death certificate dates of birth and a total of ten dis- crepancies noted. The average error for the ten discrepancies was 1.4 years. The average error for the 104 investigated cases was .35 years. The magnitude of the discrepancies ranged from three years to one year. The mean date of birth for the questionnaire data was 1884.38, and for the death certificate data 1884.23. The difference in the two means was .15 years. The standard deviation for the questionnaire dates of birth was 2.78 years and for the death certificate data 2.79 years. The investi- gated years of birth ranges from 1856 to 1912. The critical ratio was computed for the difference in the means [C.R.=.75], and the difference was not significant at the five per cent level. The correlation coefficient obtained for the two sets of data was 0.99. That the questionnaire responses investi- gated, in regard to date of birth, were valid is evident by the very high correlation between the questionnaire dates of birth and the death certificate dates of birth. 22 Validity 2: Ages 9; Death In regard to ages of death, twenty-two discrepancies were found when the questionnaires' ages of death was compared to the death certificate ages at death. The average error for the twenty-two discrepancies was 1.5 years, and the aver- age error for the 104 cases was .32 years. The errors range from three to five years. The ages at death ranged from eighty-six to twenty—eight years. The mean age of death on the questionnaires was 60.16 and on the death certificates 60.26. The standard deviation for the questionnaire data was 13.95 years and for the death certificate data 13.90 years. When the critical ratio of this difference [.10] was computed [c.R.=.125], it was found that the difference in the two means was not significant at the five per cent level. The correlation coefficient of the death certifiCate ages of death and the questionnaires ages of death was 0.98. A correlation of this magnitude indicates that the question- naire reSponse pertaining to ages of death to be valid. Validity g; the Causes pngeath Of the 104 causes of death investigated eighty-five, or 81.7 per cent of the questionnaires listed the same cause of death as the death certificates. Nineteen or 18.3 per cent of the questionnaires did not agree with the death cer- tificates in regard to cause of death. Of the questionnaire 23 responses that did not agree, two were classified as major errors, thirteen as minor errors, and four as insufficient information. The major errors constituted 10.5 per cent, the minor 68.0 per cent, and insufficient information 21.5 per cent of the total errors. The percentages of errors for the 104 cases were: [1] major errors 1.9 per cent; [2] minor errors 12.5 per cent, and [3] insufficient information 3.8 per cent. Table I illustrates the distribution of the causes of death before the investigation and after the investigation when all the errors had been corrected. The distribution on Table I indicates that most of the errors were concerned with deaths due to heart disease. The difference in the number of deaths due to senility and unknown causes [five before and two after the investigation] was brought about by specific causes of death obtained from the death certificates. Figures 1 and 2 illustrate the effect the errors have on the total distribution when the cause of death arecoarsely grouped. It will be noted that many of the errors cancel each other out. For instance, the questionnaire reSponse may have stated theczause of death as cerebral hemmorage and the death.certificate as coronary thrombosis. This procedure may have been reversed on another questionnaire and death certi- ficateEInd thus the two errors are not noted on Tables I, II, or III, or Figures 1 and 2. TABLE I 24 A COMPARISON OF CAUSES OF DEATH AS REPORTED ON THE QUESTIONNAIRES AND AS RECORDED ON THE DEATH CERTIFICATES Number of Cases Questionnaire Death Certi- Causes of Death Data ficate Data Tuberculosis of reSpiratory system 1 1 All diseases classified as infective and parasitic 1 l Malignant neoplasms, includ- ing neOplasms of lymphatic and haematOpoietic tissues 19 19 Diabetes mellitus 2 2 Vascular lesions affecting central nervous system 11 12 Chronic rheumatic heart disease 0 2 Arteriosclerotic and degen- erative heart disease 26 32 Other diseases of heart 16 8 Penumonia 2 3 Intestinal obstruction and hernia 0 l Gastritis, duodenitis, enter- itis and colitis, except diarrhoea of the newborn 1 1 Cirrhosis of liver 0 1 Senility, ill-defined and un- known causes 5 2 All other diseases 5 5 Motor vehicle accidents 4 6 All other accidents 9 7 Suicide and self-inflicted injury 2 2 TOTAL 104 104 n.nr.‘."!"'. {‘IIIK.-'7III.CII"OIIIIII I’lii ‘vl'llllll'illllbu'll. 25 ll] CANCER I VASCULA:R LESIONS OF THE CENTRAL NERVOUS SYSTEM ARTERIOSCLEROSIS 8 DEGENERATIVE HEART DISEASE DEATH ()F7 I I I SE|NILITYl 8 UNKNOWN (f) MOTOR VEHICLE LLJ I I (I) ACCIDIIENTS g ml L) OTHER MCCIDENTS SLIJICIIDE:8 I SELF- INFLICTED Illl ALL OTHER DISEASES I I I I I I‘T RI-IIEUIIMATlIC 8 ALL OTHER I I I HEART DISEASES I QUESTIONNAIRE DATA [:1 HEALTH DEPARTMENT RECORDS (TOTAL NUMBER OF CASES -IO4) IIIILIIII 02468I0|2l4 NUMBER IS I8 20 22 24 26 28 30 32 34 36 OF CASES Fig. l. A Comparison of Causes of Death as Reported on the Questionnaires and as Recorded on the Death Certificates in Semi—Coarse Categories. 26 I CANCER I VASCULAR LESIONS OF THE I I I l I l CENTRAL NERVOUS SYSTEM HEART DISEASE DEATH UL C) ALL OTHER mSEASES 09 UJ SUICHJE 8 09 I I I 3 ACCIDENTS <12 0 l SEN'UTY a I QUESTIONNAIRE DATA I E] HEALTH DEPARTMENT RECORDS UNKNOWN I I I (TOTAL NUMBER OF CASEs-IO4) I I I I I I I l I 036 I5I82I2427303336394245485I54 9|NUMBER OF CASES Fig. 2. A Comoarison of Causes of Death as Reported on - the Questionnaires and as Recorded on the Death Certificates in Coarse Categories. 27 TABLE II COMPARISON OF ALL CAUSES OF DEATH BEFORE AND AFTER CORRECTION OF CODING ERRORS Number of Cases Causes of Death Correction of Errors Before After Tuberculosis of reSpiratory system 1 l Meningococcal infections 3 0 All diseases classified as infective and parasitic 3 l Malignant neoplasms, including nec- plasms of lymphatic andihmmatopoie- tic tissues 16 19 Diabetes mellitus 3 2 Viscular lesions affecting central nervous system 5 11 Chronic rheumatic heart disease 6 l Arteriosclerotic and degenerative heart disease 21 29 Other diseases of heart 22 16 Hypertension without mention of heart 1 O Penumonia 3 4 Appendicitis l l Intestinal obstruction and hernia l O Gastritis, duodenitis, enertitis and colitis, except dirarrhoea of the newborn 0 l Cirrhosis of liver 0 O Nephritis and nephrosis l l Hyperplasia of prostate l 0 Senility, ill-defined and unknown causes 8 7 All other diseases 8 8 Motor vehicle accidents 5 5 All other accidents 9 11 Suicide and self-inflicted injury 2 2 Homicide and operations of war 3 3 TOTAL 123 123 28 TABLE III EFFECT OF CODING ERRORS WHEN CAUSES OF DEATH ARE COARSELY GROUPED Number Of Cases Causes of Death Correctiontaf Errors Before After Cancer 16 19 Vascular lesions of the central nervous system 5 11 Heart disease #3 #6 All other diseases 32 19 Suicide and accidents, homicide and War 19 21 Senility and unknown causes 8 7 TOTAL 123 123 Codinngrrors Thirty-two cases, or twenty—six per cent, of the 123 reported causes of death were found to have been incorrectly interpreted by the investigators in the original Michigan 21 State University Study. These coding errors were not Judged as to their magnitude. Tables II and III show the 21Michigan State University Study, pp, cit. 29 distribution of the causes of death before and after cor- rections had been made. Table II illustrates the effect the coding errors had on the causes of death when they are listed in Specific categories. It can be seen that most of the dif- ficulty in coding the causes of death were encountered in the categories of cancer, vascular lesions of the central nervous system, the three categories of heart diseases, and meningo- coccal infections. When the data are coarsely grouped it is evident that the coding errors distort the data much more than the errors of the questionnaire reSponses. A comparison of Tables II and III will illustrate this point. CHAPTER V SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS Summary The purpose of this study was to investigate the validity of selected factual questionnaire data from the Michigan State University Study Of Longevity and Morbidity of Former College Athletes. Also, as a subordinate purpose, the interpretation and coding of the questionnaire causes of death by the investigators in the Michigan State University Study was studied. Death certificates were used as the criteria to determine the validity of the questionnaire data. The death certificates were located for 104 of the 123 reported deaths in the Michigan State University Study. The questionnaire data pertaining to date Of birth, age Of death, and cause of death was compared to the correSponding data on the death certificates; discrepancies were then noted. After the data were Obtained, they were analyzed by conventional statistical techniques. _C_c_)nclus ions On the basis of the evidence presented in this study, the following conclusions are set forth: 31 The dates of birth as reported on 104 investigated questionnaires appear to be valid data. [Corre- lation coefficient = 0.99, critical ratio = 0.75.] The ages of death as reported on 10M investigated questionnaires appear to be valid data. [Corre- 1ation coefficient = 0.98, critical ratio = 0.13.] The causes of death as reported on the 104 inves- tigated questionnaires appear to be valid when classified in general categories. The coding errors which were found in the Michigan State University Study of Longevity and Morbidity of College Athletes contributed to an erroneous distribution of the true causes of death. The interpreting and coding Of the questionnaires causes of death distorted the data much more than the errors inherent in the questionnaire reSponses. Recommendations The following recommendations are made for future questionnaire studies concerned with the longevity and morbidity of athletes: 1. The questionnaire should include a modified list of causes of death with eXplanatory statements to clarify the different pathological conditions. Places for information such as date of death, place of death, armed service number and armed 3. 32 service unit, race. occupation or kind of busi— ness, avocation or hobbies, should be included on the questionnaire. The questionnaire data should be investigated for validity. The interpretations of the causes of death and any other data that involves interpretation should be done by one trained person to expedite standard- ization of coding techniques. BIBLIOGRAPHY BIBLIOGRAPHY Books Cuneton, Thomas K. APhysical_Fitness Appraisal and Guidance. New York: C. V. Mosby and Company, l9h7. Dublin, Louis I., et. a1. Length g£_Life. New‘Yonk: Ronald Press Company, 1949. Kelly, T. L. Scientific Method: Its Function ig Research and i3 Education. New York: The Macmillan Company, 1932. Koos, L. V. The Qpestionnair§_;g_Education. New York: The Macmillan Company, 1929. Lundberg, George A. Social Research. New York: Longmans, Green and Company, 1942. Manual 9§_the International Statistical Classification 9: EEEEBSBS: Injuries, and Causes 9: Death. Sixth rev- ision. Vol. 1. Geneva, Switzerland: World Health Organization, 1949. Office of Vital Statistics. Physicians Handbook 9p Death and Birth Registration. Washington: Government Printing Office, 1949. Whitney, Frederick L. The Elements 9: Research. New York: Prentice-Hall, Inc., 19fi7. Periodicals Anderson, William G. "Further Studies in the Longevity of Yalg Athletes," Mind and Body, 23:374-375, December, 191 . Bickert, F. W. "Einfluss, des Wettkampfmassig Betriebenen Sports Auf die Lebensdauer und Todesursache,“ Deutsche Medizinische Wochenschrift. 55:23-24, 1929. Dublin, Louis I. "Longevity of College Athletes," Harpers ‘Mgpthly Magazine, 157:230-231, 1928. 35 Dublin, Louis I. "College Honor Men Long-Lives,“ Statis- tical Bulletin 9: the Metropolitan Life Insurance Company, 13:5-6, 1932. GreenWay, James C. and Ira V. Hiscock. "Mortality Among Yale Men,” Ya1e_Alumni Weekly. 35:1086, June, 1926. Hill, Bradford A. "Cricket and Its Relation to the Duration of Life," Lancet, 2:99H, 1927. Hartley, Perercival H. and Geoffrey F. Llewellyn. "A Study of Those Who Rowed in the Oxford and Cambridge Boat Race from 1829 to 1928," British Medical Journal, 1:658. April. 1939. Knoll, W. "Welches Lebensalter Erreichen die Ruderer von 'Oxford-Cambridge'?" Medizinische Klinik, 3U:h6A-466, 1938. Karpovich, Peter V. "Longevity and Athletes," Research Quarterly, Vol. 12, (May 1951), pp. 451-455. Meylan, George L. "Harvard University Oarsmen," Harvard Graduates Magazigg, 9:355-356, March, 1904. Reed, Lowell J. and Albert G. Love. "Longevity of Army Officers in Relation to Physical Fitness," The Military Surgeion, 69:380, October, 1931. Rook, Alan. "An Investigation into the Longevity of Cam- bridge Sportsmen,“ British Medical Journal, 4865: 773-777. April. 195h. Schmid, L. "How Long the Sportsmen Live," Sport and Health. Oslo, Norway: Royal Norwegian Ministry of Education, 1952, pp. 106-107. Wakefield, Mark C. "A Study of Mortality Among the Men Who Have Played in the Indiana High School State Final Bafifietball Tournaments," Research Quarterly, 15:3—5, l9 . Whitney, F. P. "The Questionnaire Craze," Educational Review, 68:139—1A0, 1924. Unpublished Materials Hudec, A. "A Study of the Longevity and Morbidity of Foot- ball and Basketball Athletes at Michigan State Univ- ersity." Unpublished Master's thesis, Michigan State University, East Lansing, Michigan, 1956. 36 Olson, H. "A Study of Longevity and Morbidity of Track Athletes at Michigan State University." Unpublished Master's thesis, Michigan State University, East Lansing, Michigan, 195 Montoye, H. J., W. VanHuss, H. Olson, and A. Hudec. "A Study of the Longevity and Morbidity of College Athletes.” Unpublished study, Michigan State Univ~ ersity, East Lansing, Michigan. Presented at the Third Annual Meeting of the "American College of Sports Medicine," March, 1956. APPENDIES APPENDIX A 38 NATIONAL STUDY OF LONGEVITY AND MORBIDITY OF ATHLETES IN COLLEGES AND UNIVERSITIES Form A. This Form is for graduates who earned a college letter in one or more sports. (Please Fill in this Form as Completely and Accurately as Possible) Date Name of Athlete (please print) Year of Birth Weight at Graduation from College IF ATHLETE IS LIVING IF ATHLETE IS DECEASED Present address Age at death yrs. Cause of death stated on death certificate: l _. Primary Present weight _._.____..___ lbs. Secondary Present general condition of health If answer is unknown, state the generally accepted cause 0f (Check one): death Good __._..._--.._____ ._ Fair _.__._.._...__ Was death sudden _...... or fingering Poor Was he married ._..... .. or single ____..__ Married .._..__.___.. Single ____ Person entering information on this form: (Check one) Name Address - a Relationship _--_ Athletic and General Sports History of Athlete Name of Sport High College Amateur Profes- Age School Non-School sional yrs. to yrs. of age yrs. to yrs. of age yrs. to yrs. of age yrs. to yrs. of age yrs. to yrs. of age 4 yrs. to yrs. of age Activity During Adult Life, Excluding Playing Participation In Sports Include vocational and avocational activities Number of hours of physical activity (daily or almost daily) Age Vigorous Moderate , Mild yrs. to yrs. hrs. hrs hrs. yrs. to yrs. hrs. hrs hrs. vrs. to yrs. hrs. hrs. hrs. yrs. to yrs. hrs. hrs hrs- Military Service Branch of Service Age __..__ yrs. to .__....___... yrs. Physical activity involved (check): Vigorous Moderate ....___..- Mild If more than one branch of the Service, name the others and indicate the amount of physical activity involved Economic Status of Home From Early Childhood Upward Before and During College years After College Years Comments (check one) (check one) Satisfactory ____.....__ Satisfactory __i--__-___ Unsatisfactory __...____ Unsatisfactory __ (OVER) Medical History AILMENT 1. Infectious and Contagious Diseases (State age of occurrence). 2. Childhood rheumatism (State, if possible, age Growing pains ---_.. Chorea __ Rheumatic fever of occurrence of any lst attack manifestations in this 2nd attack group). 3rd attack Tonsilitis-.. Tonsils removed__ Heart defects (give as complete a diagnosis as possible, such as murmurs, enlargement, irregu- larity, heart failure, etc). 3. Hypertension (Mention complications such as strokes, com- nary thrombosis, heart failure, uremia, etc., along with age of occurrence) 7 4. Arterio Sclerosis Coronary Thrombosis __._. Indicate frequency of attacks Peripheral Vascular Disease 5. Angina Pectoris Diabetes 6. Other Diseases (mention organ or body system afiected, and age of occurrence): Smoking and Drinking Habits Use alcoholic drinks: never _.._--_._.-._ moderately excessively Use tobacco: What form How much Hereditary History , Relationship If Living If Deceased ' Age Ailment, if any Age at Death Cause of Death Paternal grandfather Paternal grandmother Maternal grandfather Maternal grandmother Father Mother Brothers Sisters (If Hypertension, Coronary Thrombosis or Diabetes present in family, please indicate) Do you think that participation in athletics is beneficial, harm- ful, or has no efiect? Please comment; if critical of program, give reasons ._-_---- l Other comments which will provide additional information on your participation or lack of participation in sports. Some examples: “I played basketball for high school dur- ing afternoons and for a club in the evenings in 1%6.” “Did not play football during junior college year on account of fracture or operation.” “Etc ” 3 fits a 823 59.5 o 9572 5.5m a. $5.2 ammmmw on R as 9% ES as. 3 00mm as .m mm mm mm mm mmmH emmfi cmmm mm mm. mm moms «mam mu mm mama mom mm em mm mm mm Hem" mmmm mm mm Hams mmmm Hm Hm mmmH New "m so am mmmH flmmm mm mm on Hmmm 02m” mm ~m mama mam _. as me Hem. asmw mm mm .mmm. mmma m. om,. mama 0mm 8 mm mm mama «Sn 3 mm mmmm Nmma 22 m4 «m Am 8mm Cm n mm mm mm 32 man m... R 2 mama 32 NW do no mmma Sm a , mm m mmma ~me h 8 2 mm em. 33 mm 4m- mama mmm .u em mm pom" HHNm mm o mama 0mm” NH aw nmmm . nmm I R a... coma mg... ma 8 82 3mm 2 MN mm Oamd Wm H 3 no «(man mowm om NN mo Sm." mmmm 8mm an R we . 8.3 3m x . 2 2. . mm Gm.“ SR 3 mm .83 83 mm 2. 2. mmmH was. fl - om mama comm. - t. 2. mm...” Home 8 . mo 83 «3 n ma em . mmmfl 9.9.. mm 2 i2 .82 mm on mmma 3m .u am mm mm mama mmom mm mm moms, Hmcm ma mm, me «am 8 a m. i. mama .3; Ron S G 82 :2 a7 2 omma 3... .— 0“ cm on mmmfi amen mm mm mmmfl mmoH mm 4m mama now .. mm .8 8 SH son 2 a m5 m8. 3 tom an . , as R. .— mm mm mm“ mm ma mm same omom ma ma me me mmma mam .. mm mm mood Nwom mm mm mama mmoa mm mm mama mm“ . 2 mm mm 9: $2 . 3 IS a am. as 2 8 mm 5 mm. mm x mm ma mmmfi omen mm mm Hmma mama mm em. mo mo .ANmH mama emu . = - i mmmm Rom mm S 8 8mm mama 22 m.” mm mm .lofll_ 2 mm mm mmmm mmom mm om mema mama mm . mm nmmn cam \44 : m: R Ema moon 3 ms 3 mmms 83 0... mm 8% J] x mm am so mmmfi mOOm mm mm mm mama ~05 mm HM . some moms me A r mm am Home meow. mm mm mmma mam m~ oN as. mama .xtumt in . mm 3 484 mama mm o: 8.2 as m... 2 0&1 lill. .. mm m a men was .2 as mm. «mm m. 83 lflll. . ms 8 2 cm 82 mos 89.. a a am. am 8 J1 Sam 8.: lfll u ms 2 82 mmom 8 S . mmm. mmm on 8 S @91th . m. a 32 88 x mm as. as ma 3 Es JflaLIlSull m 8 5 2mm 2: N... H... 82 mmm mm «a .. d “ mm mm nmma mama ma mm mmmH mam mm an a . 9.. mm mm. «mm. 8 8 z.-- 82 am «a 2. En a _ .Cmvuuu .coSugqflwwuuc seashorglquwmwxcoumflm songs: fig 3 Mm” MwMe% .eowwwmo uwmwwumnmusmwfl humus: “Mum.” “WHO swung ohm-25 engaging“; hi .30 canoe .0 onsuo gaqoc.amxum« earns Manama” sumo mason 0 cause cases he nun mammMImmuuuum o canon o.s.o an. as . naumm an.» hfifimifi 3.3333. llllllldflfiflififlflflqfldflmflfll285 a: 8.33: 8 E... 52.6 20:525.. .. 5...... t. 1.... Ex... .5... giiaafiii .fifisial m xHszmnflw no APPENDIX 0* (1) (2) (3) Name Grad{Date SeriaIIHb. (4) Age (5) Date of Death St‘f (6) Date of Birth a e **--- (7) Martial Status (8) Race (9) Occupation (lO)Kind of Business or Industry (ll) Dates of Armed Service (12) Type of Death: Natural_ Violent other (13)Cause of Death — Internal Classification Index 1. T. B. Reap. lO—Meningococcal Infect. 2-T.B.Other ll-Plague B-Syphilis lZ-Acute Poliomyelitis u—Typhoid leSmallpox __> 5-Cholera 1H~Measles __ 6-Dysentary 15- -Typhus 7-Scarlet F. &Strep.Tht._ 16~Malaria ~_“‘_ 8 Diptheria 17- Other Parst.§ Infect.D. 9- -Wh00ping CouEh_18—Cancer [reverse side] (13) Cause of Death (con't) 19- Other Neoplasm 33- Ulcer Stom. & D'od'm 20- Diabetes llellitus 34—Appendicitism 21- Anemia 35- Intest. Obst. & Hernia 22-Vasc.Les.Cent.Nerv.__m 36-Gastritis,Deudenitis,&nteritis 23-Non—Cocc.Meningitis___ & Colitis 24-Hheumatic F. 37—Cirrhosis Liver 25- Chronic Hhum. Heart D. 38-15ephritis & Nephresis 26~ Art. Scler, & Degen. H. D. 39- -Hyperplasia Prostrate 27— Other H. D. 45-Senility & Unknown Cause.___ 28-Hyperten. With H. D. 46-All Other Diseases 29- -Hyperten. With No. H. D. _47- Motor Vehicle Accid. 30- Influenza AB-All other Accidents Bl-Pneumonia h9- Suicide & Self-inflic. 32-Bronchitis 50-Homicide & War (it) Duration: Immediate ‘ Other (15) If letter-winner, what Sports? *Copy of 4 x 6 information card, front and reverse. \OCD\)O\\!\‘F'\;JNH 28. 29. 30. 31. 32. 33. 34. 35. 36. 41 APPENDIX D ABBREVIATED LIST OF 50 CAUSES FOR TABULATION OF MORTALITY* Tuberculosis of reSpiratory system Tuberculosis, other forms Syphilis and its sequelw Typhoid fever Cholera Dysentery, all forms Scarlet fever and streptococcal sore throat Diptheria Whooping cough Meningococcal infections Plague Acute poliomyelitis Smallpox Measles Typhus and other rickettsial diseases Malaria All other diseases classified as infective and parasitic Malignant neoplasms, including neoplasms of lymphatic and hesmatopoietic tissues Benign and unspecified neoplasms Diabetes mellitus Ansemias Vascular lesions affecting central nervous system Nonmeningococcal meningitis Rheumatic fever Chronic rheumatic heart disease Arteriosclerotic and degenerative heart disease Other diseases of heart Hypertension with heart disease Hypertension Without mention of heart Influenza Penumonia Bronchitis Ulcer of stomach and duodenum Appendicitis Intestinal obstruction and hernia Gastritis, duodenitis, enteritis and colitis, except diarrhoea of the newborn *Manual Lf the International Statistical Classification Lf Diseases, Injuries, and Causes Lf Death (Six Revision; Geneva, Switzerland: World Health *Organization, l9h9) Vol. 1, p. 261-62. 50. 42 APPENDIX D—-Continue§ Cirrhosis of liver Nephritis and nephrosis Hyperplasia of prostate Complications of pregnancy, childbirth and the puerperium Congenital malformations Birth injuries, postnatal aSphyxia and atelectasis Infections of the newborn Other diseases peculiar to early infancy, and im- maturity unqualified Senility without mention of psychosis, ill-defined and unknown causes All other'diseases Motor vehicle accidents All other accidents Suicide and self-inflicted injury Homicide and Operations of war “3 APPENDIX E DBaI‘ o o o o o In reference to your reply to our questionnaire form, which you kindly filled out for us, there are a few more items of information we need to complete this national long- cvity study. Would you please fill in the following information concerning the death of . . . . . . . . . Date of death Place of death city county state Occupation or business If in Armed Services, branch Service Number We wish to eXpreSB in advance our thanks to you for this information. Sincerely yours, Earl L. Mahoney ELMzcs nu APPENDIX F Director, Bureau of Records and Statistics, State Department of Public Health 631 J Street Scaramento 14, California Dear Sir: Michigan State University along with a number of other colleges and universities throughout the country has under- taken a research project of extremely great importance to present and future generations. It is a national study of longevity and morbidity of former athletes, as compared with a random sample of students were not.active in college sports. The potential value of such a study for our national health must be evident to all. Literally thousands of boys and girls, youth and adults, participate annually in a variety of vigorous competitive sports. This participation is a result of both a strong national urge and the ever increasing encourage- ment which is being applied by the schools, colleges, clubs, and other organizations throughout the land. Would your State Public Health Department be willing to cooperate to the extent of providing us with information on the following deceased persons. .Ngmg g: Deceased Date g: Death Last Knowg Address An information card has been sent along. On these cards we have checked off the following items which we hope you will be able to complete from your records: marital status,