Tobacco Documents Online

SS,~IES 4-2 HEALTH STATISTICS FRO,¥ THE If. S. ,%~ATIOgAL HEALTII SURVEY~ ~he statistical design of the Health Household-Interview Survey by staff of the $. National Bealth Survey and the Bureau of the Census U. S. DEPARTHENT OF HEALTH, EDUCATIOH, AND WELFARE Harion B, Folsom, Secretary Public Health Service Leroy E. Burney, Surgeon General Division of Public Health Hethods Wil]iz~ H. Stewart. H. D.,Chief III Washington, D. C. July 1958 Tl06653812

IU. S. NATIONAL HEALTH SURVEY Forrest E. Linder, Ph.D., Director Theodore D. Woolsey, Assistant Director Alice M. Waterhouse, M. D., Medical Advisor Walt R. Simmons. Statistical Advisor O. K. Sagen, Ph. D., Chief, Special Studies Philip S. Lawrence, Sc. D., Chief, Household Survey Analysis Margery R. Cunningham, Staff Assistant The U. S. National Health Survey is a continuing program under which the Public Health Service makes studies to determine the extent of ill- ness and disability in the population of the United States and to gather related i~ormation. It is authorized by Public Law 652, 84th Congress. (:O-OPERATION OF THE BUREAU OF THE CENSUS Under the legislation establishing the National Health Survey, the Public Health Service is authorized to use, insofar as possible, the services or facilities of other Federal, State, or private agencies. For the national household survey the Bureau of the Census designed and selected the sample, conducted the household interviews, and processed the data in accordance with specifications established by the Public Health Service. Public Health Service Publication No. 584-A2 TI0~653B13

PREFACE This report presents a description of the initial statistical design of the continuing Health Household-Interview Survey, which is a major phase of the program of the U. S. National Health Survey. The design described inthis report is that used during the period, July-December [957, which, with minor modifications, will be used throughout 1958. Except for such modifications, the design is, therefore, the basis of the sta- tistical reports being published from the household interviews conducted during this period. General requirements for the survey design were pre- pared by the Public Health Service and on the basis of these, the theoretical and operating plan of the sample was prepared by th~ staff of the Census Bureau. Although there are some important differences, the sample plan for this health survey draws heavily from designs previously developed by the Bureau of the Census for its Current Population Survey. In addition to its function as the principal designer of the survey sample plan, the Census Bureau also conducts the field interviewing, and processes the data in accordance with spec- ifications provided by the Public Health Service. Tabulation is handled on the Census Bureauts electronic computers. Final tables and published reports are planned and prepared by the Public Health Service. Principal responsibility for development of the statistical design and preparation of the text of this report was shared by William N. Hurwitz, Harold Nisselson, Walt R, Simmons, Joseph Steinberg, Joseph Waksberg, and Theodore D. Woolsey. (Messrs. Simmons and Woolsey are members of the U. S. Na- tional Health Survey staff; Messrs. Hurwitz, Nisselson, Stein- berg, and Waksberg are staff members of the Bureau of the Census.) They were assisted by numerous members of the Census Bureau staff, including especially Katherine G. Capt. Robert Ho Finch, Jr., Mary J. Jaracz, Garrie Jo Losee, and Helen M. Lucas. TI06653814

STATISTICAL DESIGN OF THE HEALTH HOUSEHOLD-INTERVIEW SURVEY 1. INTRODUCTION The program of the U. S. National Health Sur- vey is a statistical measurement of the extent of illness, disability, and related conditions of the population. This program consists of several dis- tinct but related parts. One of these is the collection of data on health through a continuing Health House- hold-Interview Survey. A second main part of the program is a series of surveys which utilize pro- cedures other than household interview as the source of data on health. A third phase of the pro- gram evaluates procedures and results and devel- ops improved techniques of measurement. The present report describes the statistical design of the ftealth Household-Interview Survey. In addition to setting forth the pattern of the Survey as it was initiated in Jdy 1957 and as it functioned in its first year of operation, the report will em- phasize two further points. One is that the house- hold interviews, while independent in a statistical sense of other surveys in the program, are but one very important component of the broader under- taking which is the U. S. National Health Survey (NHS), The second is that the "hdusehold survey constitutes an evolutionary program which may be expected to change as experience accumulates and which, at any given time, is expected to fulfill only those objectives of the National Health Survey for which it is the most appropriate vehicle. Substantive findings from the household-inter- view survey are being published by the Public Health Service in a sequence of numbered docu- ments identified as Health Statistics. Series B. Technical reports and methodological studies are issued in Series A, and include this report on sta- tistical design, Arrangement of material in the present re- port is intended to facilitate use by two different groups of readers. It is hoped that the body of the report will be of interest to and readily readable by all professional persons coneernod with health problems and those interested in research methods. Several technical appendices have been added for the benefit of statisticians, but contain material which may be informative for a wider audience. 2. BACKGROUND AND OBJECTIVES History A detailed account was given in the first pub- lication in this Series t, of the background, need for, purposes, and expected product of the U. S. National Health Survey. That story will not be du- plicated here. However, it may be helpful to recall very briefly a few highlights of the period which preceded initiation of the operating program in the middle of 1957. By 1957 it had been more than 20 years since thelast major survey to obtain comprehensive sta- tistics on diseases, injuries, and impaixments in the general population of the United States. Carried out in 1935-36, that survey was a major project in which 737,000 urban households were visited by interviewers to obtain data on morbidity, impair- ments, and health characteristics. It remains a landmark in the field. In the years since 1936 there have been a num- ber of community studies of morbidity, prominent among which are the names of Hagerstown and Baltimore, Md.; Pittsburgh, Pa.; Hunterdon County, N. J.; Kansas City, Mo.; New York Cit~,; and Cali- fornia (both San Jose and a statewide study). These T106653815

~tubie~. as well as ozcasional emferimmnts with me~.J is ~le of pro~J~g injuries [oge~r %~th rela~ i~o~adon &uch as ~e accompany~g lo~s of ~ ~om work or o~er u~ual acd~J~es. ~ Janus., 1949, ~e U. S. NatlonM Com~ee on VitM and Heal~Statisdcs was established. Sub- committees were established in ~cem~r 1949 and ~to~r 1950 to study the needs for current mor- bi~ty sta~sdcs. As a res~ of lhe~ recommenda- tions, a third S~¢o~ttee was established in Februa~ 1951 under ~e chaima~s~p of Dr. W. ~urber F~es of Jo~s HopMns Univ~rsi~, and instructed to draft a "Plan fern nation~ surveykeepinginview~e interests ofl~al ar~s." After ~reful study, t~s S~co~ittee recom- mend~ that several s~eps be taken, and in ~rtic- ~ar: "~at a condn~ng na~onal morbidity survey ~ conducted .... ks pu~ose would he ~o obtain data on the prevalence and ~cidence of disease. injuries, and impairments, on the nature and du- ration of the resulting disability, and on the amoun~ and type of mescal care received. ~e data wouid ~ obtained from a probability sample of house- holds" (page 28 of reference i). Public Law 652 and NHS Objectives In the summer of 1955, the Department of Health, Education. and Welfare proposed legisla- tion for a continuing health survey, closely paral- leling recommendations of the Subcommittee, The proposal was included in the President's recom- mendations on health matters, received bipartisan support in Congress, was enacted into Public Law 652, 84th Congress. and was signed by the Presi- dent on July 3, 1956. Later the same month appro- priations were made available for planning and pretesting during the fiscal year ending June 30, 1957. The law authorizes the Surgeon General of the Public Health Service to make continuing surveys and special studies of the population of the Urtited States to determine the extent of illness and disa- bility and related information suchas: the number, age. sex, ability to work or engage in other activi- ties, and occupation or activities of persons af- flicted with chronic or other disease or injury or handicapping condition; the type of disease or in- jury or handicapping condition of each person so afflicted; the length of time that each such person has been prevented from carrying on his occupa- tion or activities; the amounts and types of serv- ices received for or because of such conditions; and the economic and other impacts of such con- ditions. A sigmificant feat-are cf PubLic Lay; 652 is that it not e:Lly provides ~at s~.d:srantive data h_ bled, but in additicn, directs th~ Public Health S~'ice. "to develop and test hey, or improved mer/-tcd3 for chtaining current data cn illness and disabilitT and related information." Thus legislative intent Ic-~ks to the establish- menz of health statistics as noted in the law. and foresees "... continuing surveys ... spqcial stud- ies... [and] develop ling] a.d test ling] new and improvcd methods" as the objectives of the U, S, National Health Survey. Planning and Pretesting the Household fnterv'iews Throughout the fiscal year ending in June 1957 plans were developed for organizing and carrying out the household survey which had been contem- plated by the Subcommittee and authorized by Con- gress. The law contained the provision whereby the program could secure the assistance of other Fed- eral agencies, as well as private persons or agen- cies, in carrying out its responsibilities. Under this provision, the NHS made arrangements to uti- lize the very extensive resources and experience of the Bureau of the Census in planning and con- ducting the household-interview .survey. From the beginning, it was clear that the N~- tional Health Survey should be a general mttItipur- pose undertaking, rather than a study with some single specific limited objective. This concept meant that presurvey planning was particularly important. It required a careful review of previous efforts, a weighing and evaluating of a large number of possible alternatives, so that the new survey might be sufficiendy comprehensive to cover many of the desired objectives, while at the same time not to be so diluted as to deal inadequately with all topics, By February 1957, general structure of the survey had been determined, samples had been drawn, and a tentative questionnaire and field in- structions had been drafted. A pretest of 1,200 households was condt/cted in Charlotte, N, C,, to provide a complete field trial of procedure. The pretest was used alsofor training field supervisors for the national program, The next month was de- voted topolishing the questionnaireand procedures, and to hiring and training interviewers, In the 2 months of May and June, the entire nationwide or- ganization went through a shakedown and araiuing period with interviewing and editing proceeding just as though the survey were in operation. Of- ficial collection of data began the first week in July 1957. TI08653816

3. SUMMARY OF STRUCTURE OF HEALTH HOUSEHOLD-INTERVIEW SURVEY Role of Intervlew Survey. As noted in the previous .~ection, the program of the U. S. National Health Survey is intended to be an intensive and sustained undertaking to pro- vide morbidity and health statistics, utilizing what- ever resources andmethods are appropriate to the task. Tt~e program is expected further to evaluate existing sources and methods and to develop new methodologies. Among possible sources of data a prominent position goes to medical and health records. These include such originating places as hospitals, physi- cians' and dentists' offices, and insurance records of several kinds. They include, too, reporting under governmental regulation of certain types of mor- bidity and mortality, and especially the filing of death certificates. Another potentially significant source of in- formation may lie in samples of persons who are given clinical tests and measurements or general health or medical examinations. All these sources, and others, are to be ex- plored by the NHS. Several pilot projects in these areas already have been initiated. However, a considerabl~ body of opinion con- siders the household interview as one of the most promising sources of data on health. There are limitations to the accuracy of diag- nostic and other information collected in household interviews. For diagnostic information the house- hold respondent, can, at best, pass on to the inter- viewer only the information the physician has given to the family. For conditions not medicallyattended, diagnostic information is often no more than a de- scription of symptoms. However, other types of facts, such as those concerning t.he circumstances and consequences of illness or injury and the re- suiting action taken or sought by the individual, can be obtained more accurately from household members than from any other source since only the persons concerned arc in a position to report all of this type of information. Furthermore this type of survey facilitates greatly comparison of the ill population and the well population, and as- sessment of relative impacts of a variety of ill- nesses and impairments. The Health Household- Interview Survey described in this report is the vehicle being used by the U. S. National Health Survey to produce data presently believed to be most appropriately obtained from members of the household. Evolutionary Pattern Continuity and comparability of estimates for different time periods are desirable objectives, and will be given attention in the interview survey, especially when changes are proposed but they will not have overriding priority. A substantial portion of resources and energy of the NHS, at least dur- ing its early years, is to be devoted to studies and evaluation of quality of data input, to efficiency of collection and processing, and to usefulness of out- put. It is expected that these activities, augmented by the active and constructive criticism of users. will lead to a program which is changing in re- sponse to need in scope, content, method, and spe- cific product. Although the interview survey has only had 1 full year of operation, already changes have been made in sample design, questionnaire, and collec- tion and processing procedures. The description given in the follo~ving pages is in all major re- spects that which was in effect through the first year of operation, although minor changes occurred from one quarter to another. Quantitative refer- ences suchas sample sizes and noninterview rates apply for the most part specifically to experience in the first 2 quarters of operation. The Questionnaire The questionnaire is a 9-part document which is handled by the interviewer rather than the re- spondent, and on which the interviewer transcribes replies of.the respondent. Most replies can be re- corded by checking proper boxes on the form. The text of the questionnaire is supplemented by6 check list cards which are shown to the respondent at appropriate points in the interview. The check lists clarify certain questions so as to aid the respond- ent in understanding types of answers required and in recalling specific experiences. Physically, the questionnaire is of the hook type, providing separate columns for each of 7 possible members of a household. If a household contains more than 7 members, more than 1 ques- tionnaire is used. A facsimile of the questionnaire is contained in Appendix I. It is planned that items on the questionnaire may be divided into 2 groups--not separately ex- T108653817

i-/hited in the prem~ format. 0~ group efa core of ~asic quz$~ns w~ch ,~I ~e re~a~d ~n rela~vely unchanSed fo~ over an exTen~ ~od ~f tim~. ~e me~nd ~oup consists of sup- plemenm~' questions which will ke inclul~ ~em- psra~ly for blccks o[ I or a few ~len~r quar- ters. ~s general plan provides for ~e of re~ar se~es of ~sic stads~cs, and at same time ~e~its fle~flity in secu~ng oc~- sional measures of a %~der class of phenomena. As initially used, ~e questionnaire ca~ies 40 items for identification of households and persons and soci~cono~c desc~pdon of respondents. (A question to which the inter%ewer must secure an answer is interpret~ as one item in this count. ~e same inte~retation applies in ~e following counts.) It includes 12 general questions on presence or absence of illness, accidents, ~ir- ments, or con~tions for each member of the house- hold, and 54 detailed questions for each ~rson-- for whom the questions are appropriate--on de- tails of illnesses, accidents, and impairments, and on m~ical, dental, and hospital care. For most questions, the re~ll peri~ is the previous 2 weeks. But for some items of lo~v incidence, for which memo~ is reliable, such as hospitaliza- Oons, the recall exteads over the year pre~ous to the interview. Interviewin~ is conducted in the home, when- ever possible with the in~vidual person if over years of age, and otherwise with a responsible adult mem~r of the family. A separate report on the questionnaire is in preparation. It will trot more thoroughly the def- initions, concepts, scope, and content of ~e sched- ule. In addition, each report issued on a substantive h~Ith topic trots that part of the questionnaire w~ch applies most directly to ~e topic under study. Sample Design, Survey Methods, and I:stimation The sampling plan of the survey follows a high- ly stratified multistage probability design which permits a continuous sampling of the civilian pop- ulation of ~- United States. The first stage cf the de~i~n coasists o5 an area sample of 372 from among ahzut 1,9~0 geo~aphi~lly de~ned p~m~" s~ ~ts (PSU's) into whi~ the con~nen~l U~ted States has been ~Sded. A PSU is a counD', a ~oup of cen~guous counties, or a Stan~rd ropolitan Ar~. With no lo~s in general undersmn~ng, the main~g sta~es~wl~ch consist of a s~ies ~f sam- plings of successively smiler parcels of land-- can he telescoped and treated at this point in the report as an ~t~ate stage. Within PSU's then, d~mate-stage ~its called se~ents are defined, also ge~raphically, in such a manner that ~ch segment contains an expected 6 households in ~e sample. For each week a random sample of about 120 se~ents is drawn. Persons in the appro~- mately 700 households in those segments areinter- viewed conceding illnesses, injuries, chronic con- ~dons, disability, and other factors retated to health. Household members interviewed each week are an independent representative sample of the pop- ulation, ~o that samples for successive weeks can be combined into larger samples for, say, a calendar qua~er or a year. Thus, the desi~ permits both continuous measurement of characteristics of high incidence or prevalence and, through the larger consolidated samples, more detailed analysis of less common characteristics and smaller cate- gories. The national sample plan over a 12-month peri~ includes appro~mately 115,000 persons from some 36,0~ households in a~ut 6,000 segments, wi~ representation from every state. The desi~ is such that tabdaOons can ~ provid~ from the annual sample for various geographic sections of • e United States and for metropolitan, urban, and rural sectors of the Nation. Estimation is accomplished by a technique wMch insures that sample results are consistent with official Census Bureau estimates of current population by age, sex, and color, and which se- cures si~ificant reductions in sampling variance. Tec~ically, ~is pr~edure is a 2-stage ratio e~- Omation. Subsequent sections in the b~y of this report and in the Appendices descri~ leading fea- tures of the design ~n greater detail. 4. SURVEY PROCEDURE Collection of Data Data are collected through a household inter- view. Over the Nation there are 120 interviewers, trained, directed, and guided by 17 supervisors located in Census Bureau Regional Offices. The 4 supervisors are career Civil Service employees whose prime responsibility is the National Health Survey. They have administrative and clerical support from the Census Bureau field organization, and direct technical guidance from a Health Sta- tistics Branch in the Washington office of the Cen- sus Bureau. T106653818

Tn-~ inter'dewers (initially aH ~'om~) are gar~-tim~ employee~, selected through an exami- nation and ~es~n~ p~ce~s ~]~ch is a~i~stered hy ~e suF~isors, accor~n~ to ~ec~ons set ~ ?;as~ton. ~e ~c~t of work done ~y an in- ~e~;ie~er v~es deFen~ en densi~" of ~ sam- p~e n~r her home london. A ~i~l inte~iewer m~y have 26 assi~ents in a }~r, or an average of I assi~t each 2 ~eeks. Usually an assi~- meat consists of ~e~iews in appro~mately 12 households. Inclu~g trai~n~. ~avel, and call ~cks, ~he ~yp[cal ~ervle~er is employed an aver- a~e of 12 ho~s per week. Trai~n~forbo~ su~rvisors and~erviewers is a pr~ess for ~prov~ and con~ollin~ ~e inte~iew and dam from it. As such, it is a pro- cedure, par~s of which mus~ con~nue throughout ~e life of the su~ey, and is not an ac~ivky which co~d ~ completed at ~he ~i~in~ of ~e opera- dons. ~e su~rvisor is ~iven 5 ~nds of trainin~ ~yond ~e Ci~l Se~ce requirements for sppoin~ent ~o ~e job. First, the su~isor is supplied wi~ written ~ck~ro~d m~terials se~tin~ fo~h ~e objectives, and pu~eses of ~e ~dertakin~. Sim- ~arly he is ~iven demil~ inspections cove~n~ every aspec~and i[em of field operations. He s~ud- ies the materials, d~s practice exercises, sad rakes written examina~ons. ~e second bl~k of ~rainin~, for [he firs~ ~roup of supe~isors, was participation for2 weeks ih ~e dress-rehearsal pretes~ of the su~ey which t~k place in Charlotte. Replacements have simi- ]at experiences while se~in~ as understudies another supe~isor. ~e ~hird type of trainin~ comes from the condnuin~ flow of written instructions and corre- spo~ence, and of evaluations of ~ormance ou£from Washington. ~e latter come from quality- control sad q~lity-checkin~ operations pe~orm~ in Washington as psr~ of ~he edidn~ processes. Twice a year (3 dines the first year) super- ~sors over ~he Nation are assembled for a 2-day review of pr~ram objectives, new developments, and selected proc~ursl problems. ~ese sessions permit, of course, a helpful exchange of ideas amon~ supervisors sad between the field super- visors and the Washin~on staff. Finally, ~e supervisor has the advantage of condnuin~ experience since his regular job includes • e ~rsinin~ of interviewers, obse~adon of inter- ~ewiu~ for new inte~qewers, and personally re- in[erviewin~ a s~sample of households as a of ~he quali~y-conzrol pr~ram. As sta~ed above, ~he prospective in[erviewer is selected ~hrou~h a pr~ess of e written e~m- ina~on and testin~ of ~enersl ~eHi~ence and for apzRude for su~ey operations w~ch she wo~d exacted to pe~orm. ~e new in~e~iewer is @yen a 5-day initial co~se of ~ainin~. ~is course consists of 5 ~y~s of activity: (I) Instruction from a field ~z~rvisor on ~u~os-= and ge=~ral charac- teristics ~f~e s~:ey. (2)A demiled Fage-~--~age te~be~%er and su;e~-issr go ~ou~ aH ~c- d~n~ mate~al ~c~e~er. (S) ~as~reom Frac~ce Exercises. ~ %Td~ ~e ~te~ewer solves ~Ti~en proble~ and %~dth ~e ~sor de~e~es correct a~w~rs--~ese are exercises ra~r than zests, and ~y ~he inte~Jewer defi~tely how we~ she has succe~. (4) Home assi~ents which also are ~en answers problems, which are ~eat~d more in ~e ~ure of tests sad ~ w~ch res~ are ~scuss~ by inter- ~ewer s~ su~isor. (5) Practice in households ~der directperso~l obse~;adon by ~he supervison ~e s~udyofins~cdons, ~he prac- tice exercises, and ~he home assi~ents are ~s- ~rib.Zed ~rou~hout the 5-day peril. If ~e pros~cdve ~ze~iewer successfully completes ~e tr~nin~ course, she ~ins opera- tional inte~ie~in~, her firs~ assi~en~ ~rried out a~in ~der ~rec~ person~ observa- tion by ~e supervisor. After sppro~stely 1 month, a new ~er- viewer is ~v~ f~her Home Assi~me.~s which a~ain are ~aded and discussed, if neces~ry, by the supe~isor. Subsequently, in co--on wi~ inte~iewers, she spends 2 hours each month on such assi~men~s. Each quarter the supe~isor recon~acts abou~ one sixth of ~he households in his par~ of the sam- ple. He audits ~he household informa~on obtained ~rlier and reinte~ie~s indepe,dendy one predes- i~sted me~r of the household. He compares • fferences ~veen the ~ointe~iews and attempts to de~e~ine w~ch i~orms~ion is correct. ~ese rein~e~iews are randomly dis~ribut~ amon~ the in~e~iewers ~der his supervision so ~ha~ control charts based on s~ut5 percent of sn interviewer's work can ~ maintained. Each ~eek, as a par~ ~he editin~ pr~ess in Washington, error rates are calculated separately for ~ch ~te~iewer's work. ~ese are ~ansmit~ed ~o the appropriate su~r- ~sor for his use in further trainin~ and in ti~ht- e~n~ control over ~he In~ervIew process. Two or t~ee times each year, ~roups of in- reviewers are assembled et Ke~ionsl ~ices for I- or 2-~y refresher co~ses on objectives, mesh- es, proc~ures, and special ~ures of ~he sur- vey. After a household has been selected ~or the sample. ~ "Dear Fried" letter, si~ned by ~he rector, Bureau of ~he Census (fi~. I). is addressed and a few days ~fore [he exacted ~te~iew mailed to the household. This letter is intended ~o ~ s ~eneral in~ucdon ~o ~he su~ey, to have ~ect of addin~ o~cial ssnc~on to i~, and ~o make i~ somewhat ~sier for the in~e~[ewer to secure sn au~ence. ~ no precise address is ~o~, ~his 9tep is fore~one. Wan ~e Mte~ewer a~ives at the household, after s ve~, brief ~ucdon, she he~ns imme- T106653819

Form-hq-IS-6C4) (4-26-57) F~A~MF_NT (D~" COMMKI:ICt=" WASHIN~TON 25 6 D~ar Friend: The Bureau of the Census has been asked by the Public Health Service to act as its agent to carry out a survey to obtain information about illnesses, diseases and injuries among residents of this area. The survey is one part of the National Health Survey Program which Congress recently authorized because of the need for up-to-~late sta- tistics on the health of our people. Physicians, research workers, and other ~roups in health fields are much interested in the knowledge which will be gained from this survey. Every month several thousand addresses are chosen to give a cross-section of the whole United States. and the people at those ad- dresses are interviewed to obtain the necessary information. This month the address of your dwelling place is one of those chosen, and you will be visited by a Census Bureau interviewer within the next week or two. The interviewer will ask you a number of questions about the health of the members of your family, particularly about the illness and injuries you have had in recent weeks. Your cooperation in helping complete a questionnaire will be very much appreciated. The information you give will of course be held in confidence, We have the assurance of the Public Health Service that the informa- tion will be seen only by authorized personnel of the two agencies and that nothing will be published except statistical summaries in which no individuals can be identified. Sincerely yours, Robert W. Burgess Director Bureau of the Census TI0o653820

is a~ked e:,mcfly as pP~ cn ~ quez~or~e. Required ~fo~m~cnfor ~ch ~er~on ~ ~e responsible ~rEon 18 y~rs o~ age or older and a~ ~me at ~e ~me of~e ~; c~e~e bTa relat~ ~er~n who is r~arded as q~fied to We accu- rate i~oma~on. ~s ~efi~fion ~ an eli~le ~on~ is s~Hed ou~ ~ some detail in ~= ~ter- viewers' h~u~. ~ summa~, a~wers for ~en are Wen by a related ad~t; for a miss~g adult, by wife, parent, or ad~t son or daughter; or for an ad~t not related to ~= head of th~ house- hold, o~y by himself or a =elated ad~. Early ex- perience indicates that for persons over 18 years of age, 58 F~rcent are "self-respondents," w~le ~e remainder for whom another p~rson was informant are designated "pro~-respondents". ~e inte~iew averages ~ minutes. Immediately follovdngthe in~e~iewa "Tha~ You" letter si~ed by the S~geon Gener~ of ~e P~lic Health Se~- ice is hand~ to ~e respondent (fig. 2). In order to minimize ~avel ~me, wor~cads are so arranged that when an inte~ewer is in a neighboxh~ for an inte~ewing assi~ment, he carries ou~ necessa~ list~g operations for seg- ments which are in ~at same neigh~rho~ and w~ch will appear in samples,for ~e next 2 ~len- ~r quarters. Ap~n~x VI sets for~ in some rail ~e manner In which assi~en~s are random- ized over each quarter so that ~ch week's inter- ~ewing constitutes a random sample of th~ popu- lation, and within r~sonable arrangements of workload is widely diversifi~ by ge~raphy and inte~iewer. ~e following statistics for the first 6 months of operation shed added li~t on selected aspects of the collection pr~ess. ~ aH addresses initially schedu1~ for inclusion in ~e sample, 14 percent had ~come, by tim~ of ~II, wha~ are desi~at~ as ~pe B or T~ C exclusions, which are t~es of addresses which should no~ be Inte~iewed: dwelling u~ts which are demolished or which on more caref~ inspection were found to ~ outside chosen sample se~ents; households w~ch were deleted in ~e field, according ~o instructions, through s~sampling o~rations (derails on step are set for~ later in ~e report); households which were vacant; or households whose me~rs had residence elsewhere. ~ those households in which an ~te~iew shoed ~ve ~en conductS, 6 percent were no~nte~iews. ~e perc~t were re- fusals, and five percent were not inte~iewed ~- cause of all other reasons, but princi~lly ~use no one was at home after re~at~ ~Ii ~cks. In a~ut 63 ~rcen~ of households, inte~iew- ing was completed on the first visit. PerCent of households for which varies n~s of revisits proved to ~ necessary are sho~ ~ ~e following breakdown. ~u:zb=~r of visits Percent of all h~rJ~eholds All cases I 63 2 24 3 9 4 3 5 or ~ore i Editing and Processing. The interview is recorded initially in the book questionnaire, Form NHS-1. This form is reviewed for completeness and proper identification of per- son and household, hut other~vise not edited by the supervisor in the Census Regional Office. Reports are batched and transmitted to the Census Bureau in Washington for editing and fumher processing. In Washington, certain control operations are performed, reported information is ceded with special attention being given to medical coding, and to adequacy of data for medical coding (editing reports on inadequate information are returned to Regional Offices for future use in training and in- terviewer control), and the data are transcribed to document-sensed cards and then to punch cards. These cards are processed on conventional punch- card equipment mainly for purposes of interviewer control and for a more thorough check for com- pleteness of entries. Rejects are returned to clerks for review and correction. Corrections and addi- tions are punched and added to the deck. Informa- tion on cards is then transferred to magnetic tape, and further processing is handled on Univac elec- tronic computers. The computer carries out 4 basic opera- tions: (I) an edit of the raw reports; (2) the gener- ation of data from edited reports (e. g., bycounting number of chronic conditions reported for a per- son, to generate the statistic "number of chronic conditions reported for a person"); ~(3) estimation of specified statistics, including all necessary computational steps such as insertion of sampling rates and adjustment for noninterview; and (&) ar- rangement of estimates into derived statistical tables. As for any job of processing and editing re- turns in a sizeable survey, a myriad of steps is necessary. Most of these need no mention in this account. A few circumstances are worth noting. Information moves through 4 separate' chart- nels in processing, each channel being identified as a card, and each card containing the class of information indicated by its title. The four chan- nels are household cards, person cards, condition cards, and hospital cards. In nearly all surveys the choice of definitions and of categorizing devices is critical to the un- T106653821

•.. for the assistance you have given the Census Bureau interviewer who just visited you. It is only through the cooperation of you and others who are being visited that a health survey such as this one can be carried on, and we thought you would like to know how the information you have given will be used. It will, of course, be held in confidence• When combined with information given by other persons in this and other com- munities, it will reflect health conditions throughout the United States and provide new knowledge to improve the health of the American people• It isbecause such knowledge is now lacking that Congress recently authorized the National Health Survey--of which the interviewing in this area is a part. The National Health Survey will be collecting information on other aspects of health, and it is possible that we may wish to ask for your further cooperation at some time in the future. Meanwhile, thank you for your help today. Surgeon General, Public Health Service Figure 2. Letter o£ ~pprec;ation to respondent. 8 T106653822

d,-~23dn3. This statement a~plie~ vdth ~ many ways ~ass~fion dete~es ~e d ~= pro~ect, decIde~ ~'k=~=r ~m ~ a of chosen b1~:~ ~ i~o~affon, ~uenc~ ~e pre~ence or absence of bias in ~a meas~emen~ ~zess, and genera~y con~fio~ ~ ufiHW s~,ey re~ts. For ~ese r~sons, anti adequate ~ea~ent of ~ds one matter is lengthy, a separate report on defi~fio~ and class~icati~ns is ~g pr~ar~ for issue in ~s s~ies. 0~y a few remarks are includ~ here. " V~erever ~ssibl~ 9renard-~efi~dons and classifications have ~en employS. ~us, "dwel- lin~ ', u~t,'~ "household," "Standard Metropolitan Area," "family," and many other terms are de- fined as in ~e ~cennial Pop~ation Census or other widely accepted operati~s. Similarly, the Inte~ational Statistical Class~icafion of Diseases, Injuries, and Causes ~ Dea~ is ~e basis of clas- sify~g health conditions; dem~raphic, s~ia], and economic measures have been ~ouped into classes which co~orm, it is ~lieved, to most common practice. Classifications have ~en pr~esi~ated and are fixed, and the questionnaire is ~eared in most areas to such a system. Ve~ little latitude is given the respondent to create new ~asses through replies, but translating replies ~toa specific c~e is still a relatively ~fficult pr~ess, especially for ~e mescal c~in~. Accordingly, m~ic~ c~ing for each condi- tion ini~ally has ~en done independently by 2 c~ers, with differences ~ing umpir~ by a c~ing experL Information conceding the naive of c~- iug ~fficulties is ~g assembled to permit par- tial verification and pr~ess-control techniques for the medical c~ing. The use of p~ch-card equipment for early phases of editing was dictated at first by consider- ation of workloads on available eq~pment, and of start-stop requirements for some of the steps. Some of these operations have ~en shifted to ~ computer, and others will ~ later. When reportsarereceived in Washin~on, ~ey go througha control procedure ate "con~ol desk." ~is pr~edure, in addition to routine housekeep- ing checks, includes 3 operations with statistical si~ificance. i. la the prelisting step, an "expected" n~- ~r ~ households in each se~ent was dete~ined in the field and r~orted to Washington. Incoming reports must account for ~e same num~r of . households or explain ~screpa,cies. Any segment for which an unexplained discrepancy is found reconciled through reconmct wi~ ~e Census Bu- r~u Regional ~fice if time ~rmits. In instances in which tab~ation cutoff ~me prevents this, the case is later call~ to ~e atten~on of the Regional ~ice so that it will ~ used in initiating neces- sa~ ti~tening of su~rvisory controls on listing, inte~ie%~ing, and cleric~ operations. 2. In s~me s_~grn~nts it will ,have keen fo-.md, eid-.er earlier in Washington or in ~e field, a chosen se~n-nent clearly centa~-~d more than 20 households. In these cases the Ee~'nent was ~ampled so tl--~t the final ~l~sample canLained roughly 6 h~useholds. Tile s~tbzampling fraction is noted at the Washington centrol desk and an ad- justing order is g~ansmitted to the computer. 3. A third D~e of review at the control desk adjusts the sampling fractJoll for households and persons from special dwelling places, such as re- formatories, homes for the aged, or hotels for transients. The general purpose of all operations at the Wasbdngton control desk is to assure that dam moving into the editing and tabulating stream are, with respect to coverage and weighting, in agree- ment with the survey design, within narrow toler- ances of error, Evaluation and Control of Data A substantial proportion of total resources of the household-interview survey is devoted to con- trol, evaluation, and improvement of quality of data input. There are 4 very broad areas of activity which have impact on quality of data, which are not discussed here in this connection, but which are pal~s of the Uo So National Health Survey, and which are listed by tide in order to place in perspective those items which are displayed in the following paragraphs as devices for control and improve- ment of data. The 4 areas are (I) the over-all survey design, including concepts, definitions, and general plan of operation; (2) operating control, in the sense of maintaining general adherence to design, including proper use of trainin~ and super- vision; (3) utilization of comparative analysis of data, including external checks against other sources of health information, and especially against medical and health records; and (4)pro- vision for organized outside review and criticism of both methods and products through the creation of both governmental and nongovernmental advi- sory committees, and the use of consultants. Aside, then, from these broad areas just named, there are 3 vlpes of operation which are integral par~s of survey procedure, and which are principally devices for control and improvement of quality of data. The Keinterview Procedure.--Already men- cloned, in connection with training and supervision of interviewers, is the reinterview program, The supervisor regularly recontacts about one sixth of the households in his part of the sample, and thus, about one sixth of the households assigned to each interviewer. The supervisor audits the household information previously secured by the original interviewer, and reinterviews I pre- designated member of fine household. Three main 9 T106653S23

ef ~imng an~ q~i~' con~olfo~ bili~'. ~ ~d is de~ecfi~ of ~t~ewer m ~ extent ~t the more e~en s~-~or c~ ~scover it. ~s m~ns e~ra~g ~e as~ption • at ~a su~e~i~or, us~g ~e s~e ~estio~e ~d ~ a~e p~ed~e ~a~ were ~ by ~e ~i- ~al ~te~ewe~, hut~more~orou~y ~ainad, will s~ure ~ta w~ch may ~ consid~ ~mn~rd w~ ~e ~te~iewer ~ho~d and pass~ly ~d. ~ ~e reinte~ew, ~ ad~ mus~ ~ inte~iewed as se~-res~ndents ~an proxy-respondents. ~us ~ere is a com~on~nt of variance from se~-~, pro~-~e~d~t as source. This componem has some ~ufing on measurement of inte~ew~ con~butions bias and variance, but its e~ate~ce ~y make possible ~o de~erm~e the exten~ d bias (ff any) caused by the proxy-respondent. Pr~esa~g checks and con~ds.~t ~ch c/pal p~easing step, c~trols are established either by verification ~ough duplicate or by sample verificar~on te~niquez, bas~l pn- ~y en prc~ezs con~oL ~ far, ~or rates ~yanl prelimina~, s~n~rds ~e ~ed to ~ at- tendc~ of ~e res~ibl~ ~;~radn~ su~sor. ~ a reco~on ~at steps ~e token ~o re- du~e th~ e~or ~te. F~er s~y of e~or ~[es and ~e~ pro~le impact en es~mates ~e ex- pect~ to I~d to a he~er-~anced se~ of stand- ards. ~ney will l~d also ~o ~t~r use ~ s~ple ver~on and reduc~ons in I~ percenl dupli- ca~ of e~fing steps. ~te~al ~fin~ and consistency checks.~efer- en~e was made ~r~er, and will ~ made aga~ when the derails of es~afion in the su~ey are ~scussed, to e~ng routines desired to make questionnaires in~e~a~y consistent, and to e~m- i~e "impossible" res~nses. ~s is an ar~ ~ w~ch the m~ber and ty~ of poss~le checks are ~mited. E~erience must ~ ~e ~ide in de- ciding how much e~ting is profit~le. As ~plied earlier, the firs~ objec~ves are to ~s~e ~ha~ data are consistent and not obviously incorrect. More p~ne~rat~ e~ts are to ~ tes~. 5. SAMPLE DESIGN :Th'e Multistage Design. As noted in the summax~] on page 4 of this report, the Health Household-Interview Survey rests on a highly stratified, constructively 2-stage probability design. Acmalselectton of sample units takes place in a multistage process, which is mod- ified further by the use of 3 selection zones and 41 subuniverses. The design is termed "construc- tively 2-stage" because the first sampling step is the selection of 372 primary sampling units from among some 1,900 areas into which the country has been divided, while the remaining steps lead effectively to a second or ultimate sampling stage in %vhich small segments or clusters of an expected 6 households are chosen for inclusion in the sam- ple from within the PSU's selected in the first step. The following paragraphs describe principal features of the design, and the manner in which the sample was drawn. Additional technical notes on selected aspects of the design are included in Ap- pendices [I through VII. In particular, algebraic statements of the estimating and variance equa- tions are given in Appendices II and IlL Still fur- ther insight on the topic can be gained from con- sulting Chapters 7, 8, 9, and 12 and AppendLx B of reference 2, since much of the theory underlying the sample design of the health survey is set forth in this b~ok. I0 Primary sampling units. The PSU is a county, a group of contiguous counties, or a Standard Met- tropolitan Area. A total of 1,900 PSU's exhaust the land area of the continental United States, Forma- tion of such PSU's is an art rather than a science, although several clear-cut principles and rules were helpful. Prominent among these are the fol- lowing 4: 1. PSU's should be units for which a wide va- riety of descriptive statistics is I available, since this permits the PSU's to be stratified or classi- fled in an efficient manner. 2. When the PSU is used by a large surveying organization, there are distinct economies in using the same set of PSU's for more than 1 survey. Consequently there are advantages in having the PSU conform to administrative structure in the field, and in having the unit adaptable to many so- cial andeconomic objectives. 3. For technical sampling reasons, the great- er the internal heterogeneity of the PSU, the more efficient it is. This principle tends to produce physically large units. 4. Contrastingly, costs per ultimate sample unit (i. e., cluster of sample households) tend to increase with transportation distances between ul- timate units within a PSU, and thus to increase with the size of the PSU. This factor has limited the size of a PSU to not more than a few neigh- boring counties, _J T106653824

Th~ aho;-e pzinaiples led to formation of I,g69 PSU's, w~ch are ~ss ~d ~ oi~r E~u ~'eyz, and ~'~ch, ~i~ a fe'~' ~c~p~o~, have ~e f~e~: ~ h~n~ bl~k or ez~ s~c~al com~n~nt of ~ PSU ~ a co.W; ~ch PSU ~ ~ W~e~ U~ed Smt~ contains a ;~p~adon ~f at least 7.5~ (19~0 Ce~us). and o~r parts ~ ~ co~ a pop.salon of at l~sz 10,~O; ~ch wes~e~ PSU conta~s not more ~ 2,0~ sq~re ~les and o~her PSU's no~ more ~an 1,500 square mfles--u~ess ~e single county is larger, which in ~e West resuh~ in many PSU's hav~g less ~n 7,5~ persons; and, ~ith ~he q~- ~cadon ~ each Stan~rd Metropolitan Ar~ a PSU. ~e PSU is kep~ as in~e~aHy contrasting as possible in s~i~conomic ~ems. S~dficad~ of PSU's~ampling ~eo~ makes it cl~r ~at if units to ~ ~mpled ~n ~ classi- fied into ~ego~es or strata whose mem~rs tend m ~ relatively ~ke wi~Nn strata and r~advely u~ike ~ween s~ata, and drawings ~de ~om • ose s~am. ~hen resulting sampling variances are z~uced over those of samples ~awn from an ~stratifi~ universe. ~e PSU's were s~at~ied accordingly, ~e principal m~es of stradficadon ~ing ge~rap~ l~a~ion, densiw of popula~on, rate of pop~a~ion ~ow~h ~ween 1940 and 1950, pr~o~on d n~whi~e, type of ~dus~ in pre- domi~nfly ur~n areas, and ~ of farming in ~ral areas. The general sampling desi~ con- ~emplat~ drawing f~st-s~age uni~ wi~h pro~bil- i~y pro~nio~te ~o size, wi~ 1 PSU to ~ ~a~ from ~ch s~ra~. F~zher, iz was desir~ ~hat separate estimates ~ obtainable r~dily for ~ch of 41 subu~verses--to ~ f~ther descried later, but characmriz~ often in [he survey as Tab Ar~s. ~ese s~cffi~ons, au~en~ed by an e~sGng s~raGficadon of ~e PSU's. set up by ~e Census Bur~u for o~er pu~oses, res~t~ in classifica- tion of ~he appro~mamly 1,9~ PSU's into 372 strata. Fur~er descripdon of ~e precise manner ia which ~s was done is Wen in Appen~x IV. ~aw~ first-stage u~ts~From ~cb of the $72 strata I PSU was selec~ for inclusion in ~he sample wi~ pro~bi~W propor~ioaam ~o its pop.salon. ~is m~nt. for ex~ple, that a s~ll PSU with 50,~ i~bimnts in 1950 had o~y 1/20 as much chance of inclusion in [he sample as did ~e 15rger PSU w{~ 1 million in~bi~an~s. ~&se ~eren~ial sampling ~ms were of course taken ~to consideration in subsequent sampling and es- ti~ting s~eps. As indicamd, the selection procure and s~c~icatioa ~a~ serrate worksheet est~ms ~ compu~ for ~ch of ~he Tab ~eas had i~u- enced s~a~ffi~tion. ~e Tab Ar~s ini~lly s~c- ffied were ~e 8 larges~ Standard Me~opolima Ar~s. and wi~ ~ch of 11 ge~raphic sections. ~e S s~sections composed of (1)smaller Stand- ard Metr~olimn Ar~s. (2) o~er ur~n ar~s. and (3) o~er ~ ar~s. ~e s~tions and 8 la~e S~'s are sho~m on ~e map in fig~e 3. It shoed he understc.--4 ~.at s~parate s~adsdcs ~hed for e~ of ~ ~ T~ #~s, hut ~er ~ ~ for Tab ~s ~n ~ more ~n one ~my in[o breeder ~e$o~es for which reH~le ~es ~ ~e pr~uced. In ~ome ~s~nces, effici~ s~ca~on re- s~t~ in a s~atum he~g compos~ of 1 sin~e large PSU. From such a stra~ ~e sin~e PSU enters ~e s~ple ~ cerm~D' and is ~lled a se~-representing PSU, Each of ~e 8 largest S~'s and 102 o~er PSU's ~me self-representing PSU's. Each PSU ~a~ into ~e sample from a nonme~opolitan strat~ was utilized later as the ~ame for both "o~er urban" and "other ~ral" mb~ation areas. Table 12 ~ Appen~x VIII shows ~e geo~ap~e dis~ibudon of ~ se~-represent- ing ann sonnet-representing PSU's. Selection z~es~For s~pling p~poses and in order to reduce over-all variance, ~e civilian pop~aOon in ~e United States is ~vid~ into 3 mutu~ly exclusive classes or selection zones: Zone A. ~ose ~rsons living in co~on ,dwelling places. ~ne B. ~ose ~rsons living in ar~s of "new h~sing." Zone C. ~ose persons living in large spec- ial dwe~ng places. Common dweH~g places include whatwo~d be or- ~narHy regarded as such--for example, private homes, a~rtment houses, and duplexes. Areas of new hous~g are simply ~ose in wMch considerable new housing has ~en built since ~e last population e~s~ (Apr~ 1950) and wMch have ~en recorded and mapped by ~e ~nsus Bureau. ~ese may elude ar~s which wo~d ~ class~ as ~lon~ng to eider ~ne A or Zone C except ~t they are positively identified as ~ing in ~ne B, Special" dwelling places include such places as ~tentia- ties, refomatories, homes for ~e aged, mental hospimls, and hotels for ~ansients, ~e 372 first-stage ~ts are identic~ for aH 3 zones, but later-stage samplMg is han~ed sepa- rately for each zone. For the large s~ial dwdling places, Zone C. lists of in~vidu~l institutions and organi~tions • e ~mple PSU's w~e assembled from a varie~ of sources. These list~ places are exelud~ from further ar~ sampling. S~cial inst~ctions for ~aw~g s~ples of ~rs~s from ~ne ~ are pre- par~ for ~e ~fferent ~s of s~etal dwelling places. Such ~rsons, constituting about 2 percent of ~e universe, have not ~en included ~n i~dal tab~ati~ of dam and are not ~scuss~ f~er in ~is accost. ~ne relatio~hip ~tween selecd~ ~nes A and B and ~een ~nes B and C is slightly more complex and makes ~e of t~e p~eiple ofstrat~i- ~fion aft~ ~p~ng3 and ~age ~8 ~ reference 2. ~e of ~e ris~ of ~ sampl~g, when using ~m on nmMr of households for a p~or y~r as the ~sis for selecOon, lies ~ ~e e~stence of large u~ts ~new cons~ucfion built since the prior II TI06653825

TABULATION AREAS: NATIONAL HEALTH SURVEY Standard Metropolitan 1950 Pop.. Are--~s in I,O00'S b in

;~u~ unless corrective action is ~aken will in:tease varia=ze. From the National. Hous[n~ ~e C~n~ B~u had av~a~le a r~cord of large new con~c~cn activities ~ many of ~ 572 sam- pl~ PSU's. Consider ~s~ PSU's as he~n~ strad- fi~ ~nto 2 classes: (I) ~hos= PSU's which contain ar~s of new hous~n~ so and (2)all o~er PSU's. SeEmen~s are from wi~ ~I sample PSU's ~o represent Zones A and C. ~nes A and C are mutually exclusive; ~at ~s, they do no~ overlap. ~e semen,s se]ec~ ~om ~e class i PSU's are ~hen e~mined to see ~ ~ey fa~ into areas classified as new construc- tion ar~s accor~g to ~e Housing Inventory. ~ose segments from class I which do not f~l into new cons~ction ar~s and all accents select~ from class 2 PSU's are retained and ~come the independen~ samples for Zones A and C. ~e seg- ments or parts of segments w~ch are contained in • e areas of new construction, and which were ini- tially drawn into ~e sample, are at t~s point de- leted from ~e original ~mple. An independent sample is then taken ~om among the new con- struction areas of Zone B at the same sampling rate as Zone A. ~er-all sampling ratios for all 3 se- lec~on Zones A, B, and C are identical within each Tab ~, Appro~mately 8 percent of the popula- tion and of the sample are accounted for by Zone B. Selection of so.eats in Zone A.--~us by far the major part (90 ~rcent) of ~e sample is found in Zone A. For many pu~oses, it is convenient to think of ~e sample as co, sisting only of Zone A. An outline is given here of the way in w~ch sam- pling within PSU's is carried out for selection Zone A. An example of the pr~ess is given in Appendix VI. ~e ~timate sampling ~it within ~e PSU is called a accent. It is a ge~raphically defined ~rcel w~ch contains an expected 6 households. Se~ents to ~ included in the sample are chosen separately for each Tab Ar~ in a series of steps or stages.. S~vey specifications resulted in a require- ment ~at over a peri~ of a y~r 144 segments are to ~ su~eyed in each Tab Ar~. Wi~in cho- sen segments, all households are interviewed. (As noted in Section &, if it develops that a selected segment con~ins obviously more than 20 house- holds, it is subsampl~ and appro~mately 6 house- holds in it are inte~iewed.) ~e selecOon pr~edure all.ares the num~r of segments to ~ interview~ to first-stage units in the Tab Ar~ in proportion to the size of the strat~ they represent. Se~ents are drawn with- in PSU's through a sequence of selection of suc- cessively smaller ~ts of area un~l finally a unit containing ~e e~ected 6 households is secure. ~is becomes the dtimate sampling unit. An illus- traticn of d;.~ prc.:ed,lre is ~iven in Ap;en~ VL S~ples for ~ y~r, qua~er, and ~veek.--~- fial ~amp~n~ is ~ ~n in a %ray ~~ch nhe ~e~ents repD~ed for ~ch ~en~r q~ner an ~i~nden~ ~mple of the land ar~ of d~ U~ Sm~es. ~ qua~erly samples are ad~ve and ~us ~e annual sample is 4 dines ~e size of quarterly samples. ~e samples are also random- ~ed by necks %%~n each quarter, so ~ each week's in~e~,iews become a random sample of ~he population and the wee~y samples are ad~tive within the qua~er. ~e dezail by which ~is is complished is iHuszrated in Append~ VII. ~e f~l s~ey desi~ is off.tire over each quarter. wee~y samples are unbiased bu~ n~es~riIy fol- low a more res~ic~ed design, on ~he average de- pendin~ upon s f~s~-s~a~e selection of 60 rather than 372 PSU's. Mapping of Segments and Listing of Households For each segment in the" sample, the inter- viewer is furnished 2 maps: a ~ and a Seg.___- merit Map. The Key Map shows the general location of the segment and may be a county highway map or a city street or block map. The se~nent number and approximate location of the segment (shown by the large dot beneath "Hillcrest Avenue" in figure 4 are entered on the Key Map. SCAt.E . MT. ALTO, RD. Fi~ure 4. Key Nap showin~ Segment 0534. 13 TI06653827

seme cas~s n~) s~rucn~res are shs"a,'n on uhe map. Two illus~rations cf S-~grnent Maps are aho',vn in figures 5 and 6. T'n~ s-~gment b~,mndaries, in any case, are o'itlined on d~._~ map. E___S_T.~ 0534 MT. ALTO ROAD N Figure 5. One type of Se~men~ Map. 14 N. CUSHING DRIVE 2100 2102 2104 2106 D A A A A [] APTS [] [] ~ S ~ D 2 2101 2103 2105 2107 2109. 2111 FOURTH STREET Figure 6. A second type of Segment Map. D - D~elllng F - Flat S - Store Apts. - Apartments A - Garage The numbers inside indicate the number of floors ~n the structure and the numbers a/on~ TI06653828

Intervie-,~rs are mstrncted to list all ~s~s or c~ de~c~cns of ~ places ~o#le live or mi~h[ live, incl~n5 su~ places as OrPhan' house d~veHin~s, a#~ents, duplexes, ~a~ers, tents, bonehead, conve~ hexes, and rented r~oms, ~¢lu~g evening ~'~ch Hes in- side • e de~n~ se~ent." ~e ins~cGon is sup- po~ed and ampl~ied by ~e maps and a 93-page ~dexed lisGng m~u~. Th~ Hs~g operaNon is con- ducted at a ~e prior ¢o ~terviewing, ~hus pro- viding 2 checks on coverage: one at ~e ~e of list- ing and a second at the time of interviewing. Summa~ of ~its~veral differen~ ~nds of u~s and ca~ories are menNuned in ~is section and in the App~n~ces. It may ~ helpful to reca- pillage in caps~e form ~e principal elements of teminol~,. Tab Ar~ - ~e of 41 s~universes, defined by geographic boundaries and by size and density of population. PSU - Primary sampling units consist- ing of I or a group of contig- uous counties: about 1,900 of them in the United States; 372 in the NHS sample, Strata -372 socioeconomic classes into which the' PSU's are grouped. E__D - Enumeration. District, .a geo- graphic subdivision of a PSU, usually containing between 50 and 1,000 households. ~- A subdivision of an ED containing an expected 6 households. This is normally the uhimate sam- piing unit in the survey. Selection Zones - Strata in a different dimension, based upon type of dwelling unit, and utilized in reducing vari- ance. .Dwelling Unit - Place where persons live or might live. This is the unit list- ed for subsequent interviewing purposes. Elementary Unit- There are 4 elementary units or channels for processing in- formation which are utilized in the survey: (1) the household; (2) t_he person; (3) the health con- dition--illness, injury, chronic condition, or impairment; and (4) the episede of hospitaliza- tion. The Estimating Process Some aspects of the estimating process were treated in Section 4 under Editing and Processing, and other aspects are influenced, of course, by the ~-~ple design ~ch has just ]~e~a discussed. In ",',hat follows in t/:_~ pre=~.~nt ~ecdcn, ks f~_--us of a~tention is on the estimating problem as such. Th.~ exfimattcn pro:ess in the h~dth su~:ey is b-nsically simple, although actual prccedure in- clu!es a ccnsid~able numl:~r of steps. Leading reasons for the ap#arent comptexiW are 4 in num- ter, growing largely eut of the fact that the survey pre-~uces a variety of estimates in several dimen- sions: Geographic scope.--The survey yields work- sheet figures for the Nation as a whole, and also for constituent Tab Areas. The Tab Areas can be combined into geographic divisions of the country, or into classes ~hich reflect size and density of the population in the community. Type of statistic.--Three variations may be distinguished under this heading. (I)The number or proportion of persons in the population with a specified characteristic, such as having I or more chronic conditions, or not having visited a physi- cian within the year immediately previous to the week of interview. (2)Estimated volumes of events arising from, tabulating answers to such direct questions as, "How many days were you in the hos- pital, not counting the day you left?", with editing converting the reply to "Number of hospital days in past year." (3) The incidence of a particular disease or health condition, bulk up from cumulat- ing occurrences over 2-week periods as reported by persons interviewed in successive weeks. The first type of statistic named above will be recognized as an instance of binomial estimation (modified of course by the structure of the sample design), since each individual respondent will either have or not have the specified characteristic. The second type of statistic is like the first except that the population measures involved are quantitative rather than qualitative variables, and consequently estimation is not binomial. The distinction between the second and third types of statistic is sharpened perhaps with an example. Approximately 115,000 persons are interviewed each year in the health survey, about 2,200 each week. Each of the 115,000 persons, in effect, gives the interviewer the num- ber of days he spent in the hospital in the previous year, and thus provides data which per~nit an esti- mate of the number of days of hospitalization ex- perienced by living persons in the year previous to the week of the interview. This is a type (2) esti- mate. Similarly, each week approximately 2,200 persons report their days of hospitalization in the previous 2 weeks. Summing these reports over 52 weeks of interviewing and taking account of the l- week overlap in reference periods for adjacent weeks of interviewing would provide the basis for a second estimate of a year's hospitalization, this time the resulting statistic being of type (3). More is said later on the procedure whereby estimates of type (3) are produced. Some might also wish to distinguish, under this tide, between estimates of an aggregate, such as total number of physician visits for a specified ~5 T[08653829

~e~ew$ ~ ~o~adon ~;er aweek, a q~rter, ~ ye~, or o~er ~e ~e~-~s. ~s~ r~erence per~s for ~c~ence or vol~ of even~ ~n ~e v~ed widely, mn~g for some items from a week to any m~dpl~ d wee~ ~ ~e histoD, of ~e s~ey. Fo~ of es~ate.--Tne sm~s~cs prc4uc~ ~om ~e s~a~i~ desi~ ~ough 2 stages of ra- ~o es~afion are ~e pr~ucts ofa desi~ which is mu~h more efficiem ~an a simpl~ ~ndom sample would have ~en. hu~ which necessary require sdmewhat more elaborate computation. St~ps in estimationrln ~e interest of brining out main ~r~ds of the es~ma~g s~o~. obscured as litde as possible by ~e crosscurrents jus~ not~. ~e reminder of this section is written mostly around ~e pr~ucdon of est~ates of an average n~ber of persons with a s~cffi~ characteristic. • e average ~ing ~sed on inle~iew~g over 13 weeks. ~e pop~adon referred to is the civilian noninstitudonal pop~afion of the continental United States rather than that of one of ~e T~ Ar~s. An aggregate ra~er ~n a proportion or ra~e is • e statistic ~der obse~ation. ~casional varia- tions from ~is patte~ will ~ necessa~. As in, cared cartier, incoming reports are ~as~ through controls to insure that the data in- put to the computers is consistent with sample de- sign. pro~r~y c~. and ~pable of ~ing tabu- lated. A series of mechanical edits are carried out on the computers. These edits make the question- naire internally consistent, and adjust or account for item nonresponse. Step 3 -'~toeach record of an elementary unit(person, household, condition, and hospitalization) basic sampling inflation factors are inserted. This step takes account of all stages of sampling. The factor is the reciprocal of the combined sampling frac- tion which for a quarterly tabulation varies among Tab Areas from about 1-in-2,000 persons to about 1-in-t9,000 persons. [Sampling fractions for an- nual samples are one quarter of these numbers.] Step 4 "--~atistical theory demons~ates that a "ratio estimate" for any statistic is superior to an or- dinary "inflation estimate" if there is correlation bet~veen the numerator and the denominator of the ratio. Specifically, if Y' and X' are ordinary in- 16 flation estimates cf 2 characteris-dcs cf a ri~n, y and X, re~q:~ct/vely, and if fit.~ "true" tcLal X is tmo',vn in_~_~.r.Sendy, d~_=n the ratio e~rimate = .~ X is a l:_~tter e&rimate of Y than is provided there is correlation teL-.*;een Y' ~d In this form of estimate, th_~ quantity ~-~ becomes a calibration factor for the survey. This principle is utilized at 2 stages in the NHS. In the first instance it is used to reduce sam- pling variance between PSU's. Estimates of the 1950 population which would have been obtained from a complete enumeration of the 372 PSU's hut not other PSU's in the country were compared with official 1950 population counts for each of 120 color-residence classes. Resulting factors are shown in table A. In calculation, these factors are used in the following manner, the arithmetic being carried out automatically by the computers. Consider a eample record for a person who is white and who comes from an urban nonself-representing Standard Met- ropolitan Area in Geographic Region 1. All sample records for this person are multiplied by the fac- tor 1.075380. (See 1st line, 2d column of table A.) This brings the sample data into closer conform- ity with population controls for the universe, intro- duces only trivial, if any, bias into the estimate, and reduces sampling variance. NOTE: Steps 1 through 4 are carried out wee'kly, and provide a "deck of cards" (Uni- vac tape) of edited and adjusted sample data for each week of the 13 weeks of the quarter. The "scale" of data at this point is therefore 1/13th of universe totals. Weekly data are merged later into quarterly totals. Steps 5 and later apply tothe merged quarterly decks. Step 5 --]T~espite intensive follow-up efforts, reports on some households in the sample have not been received at the tabulation cutoff. In the first 2 quarters of operation the noninterview rate was 6 percent--1 percent refusal, and the rest for all other reasons, such as no one at home after re- peated call backs. For a sample household for which no interview is obtained, any estimating pro- cedure must necessarily impute values for each statistic for which measurement had been intended. Adjustment for noninterviews in the health survey is accomplished by a calculation which as- sumes that respondents within a particular seg- ment for a quarter represent the nonrespondents in that segment. In the rare instance in which less than half of a segment is interviewed, the nonin- terview adjustment is mcdified by evidence from T106653830

Table A. First-stage ratio estate fa=tors f~-r n~nself-repreEentimg* FgU's by residenae, col~, and section ~;onself-representing SMA' s geographic section I ..... I0 ..... ii ..... Urban ~rnite l~cn%~nite 1.0753~0 1.753515 .973243 .809129 .755966 .7Z4533 1.328674 .927637 1.210693 .509411 none • 973222 .869424 1.076720 1.027738 1.179743 1.179743 • 873733 .492896 none R~ral n~nfarm Rural farm '~hite .579098 1.345792 .723622 1.580769 1.640072 Nonself-representing no~n- SMA's geographic section I ..... 10 ..... ii ..... 1.000096 1.050026 1.101374 1.022545 1.109184 1.000335 .988710 .984943 .980173 1.043498 1.020053 .791543 .891068 1.476958 1.142936 1.649665 1.065752 1.044770 1.028117 .977126 .979703 1.203531 Non~hite .678533 .673733 1.048442 .927872 .664274 .633703 1.158533 1.158533 1.461507 .428340 .889450 .715170 .899394 1.349170 1.094866 1.874840 2.502347 2.502347 1.084175 1.243901 .883253 .912743 .873888 1.276623 1.027072 1.026288 1.013235 1.108361 .807592 .870792 .776500 1.250069 .873241 1.066578 .959741 .962977 .953182 .873888 1.011195 1.086449 1.016688 1.005968 1.005739 .978919 .991135 .991556 .919709 .997690 1.000277 1.013374 1.008986 1.004002 1.016431 .890752 1.006944 Non~-hite .488372 .721580 ,695719 1.580769 .383598 .745692 .991501 1.250069 .873241 1,003409 1.301490 .728575 ,871958 1.673347 1o156285 1.071351 .968411 .990630 1.199575 1.009057 IFirst-stage ratio estimate factors for each of 8 large sPparate ~bol.tlon or~as and for the self-repre- senting PSU's is 1,000000, reports over the entire Tab Area. An illustration of the process is given fora hypothetical Tab Area: House- holds Segment sched- number uled for inter- view i 6 2 6 3 8 4 4 Tab Area total 220 House- holds not inter- viewed 0 i 0 3 i0 Segment Excess adjust- non- ment inter- lacier views 1.0000 0 1.2000 0 1.0000 0 2.0000 2 2 Data for the 5 reports in segment 2 are multiplied by the factor 1.2000 sothat the 5 reports represent the 6 households intended for interview in the seg- ment. Segment 4 in the example is of the unusual type (where less than half the households in the segment were interviewed) which leads to a fur- r.her adjustment at the Tab Area level after a pre- liminary one has been made at the segment level. The Tab Area adjustment factor is the ratio of total households scheduled for interview to total households scheduled for interview less the "ex- cess" noninterviews; that is, the factor is 220/218 or 1.0092, in the example. Data for all reporting households in the Tab Area are multiplied by this factor to account for the 2-household "excess" of noninterviews. Advantages of the ratio-estimating process are exploited further by the introduction of a sec- ond calibrating or raNo factor which brings the 17 T108653831

ezr.imates of ths U. S. ~,,~lar.ion dsrived from d~te~d consols for 76 ag~-z~>:-coIor cl~s~ a~on ff.~e factors ~g~ ~om 0.61 to 1.~6, ~S estates for 62 ~ ~e 76 closes co~ng wit~n 12 Fercent of ~e controls. Tn~ over-~l ~S e~mate of ~ U. S. p~pula~on h~ore ~s f~al adjus~ent was ~ 0.3 ~ 1 ~rcent of ~e con~ol on ~a ~s effect of ~ese 6 steps ~ (i)to ~e ~e household su~'ey ag an ins~ent for obtaining ~rcent ~st~hu~ons'of ~e pop~ation by characte~s~cs of ~ess and he~ con~ons, and (2)to pr~uce est~ates of to~ n~rs of ~rsons in the population wi~e~e s~cified c~r- acteris~cs by m~tipl~g the derived ~rcent ~s- tribu~on by ~p~aflon c~ols. ~ates are calcu- lated by obm~ng ra~os of ~e appropriate es~u mat~ a~regates. T~ations of items o~er ~an average ~r of persons ~ s~cifi~ characteristics over a qua~er are obtained in a s~flar manner, but with varia~ons in pr~e, ~e ~icular va~a- ~one de~nd~g on ~e nat.s ~ ~e item. Two ex- amples may suggest the ~ndg ~ variations w~ch are ne~ed. Consider again the ~e (3) es~mate discussed a~ve, in w~ch ~e o~ective is to obtain an esti- mate ~ ~e mini nu~r of ~ye of hospiml~a~on over a y~r, and consider first an estimate over quarter. An item on the questio~a~e asks each ~rson for the n~r of such ~ye in the 2-week peri~ i~e~ately preceding the ~len~r week of inte~iew. Each week's inte~iewing, since it an inde~ndent sample of the popMation, pr~uces, by the process descried in the 6 steps a~ve, an estimate of 1/13~ of the total hospital ~ys over a 2-week peril. (It will ~ recalled that the sampling fractions have ~en expressed inters of 13 wee~ of inte~iewing, and ~e weighting factors have set aceor~y ~ ~e computer.} M~tiplica~on by 6.5 yields 1/13th of ~e total visits for a 13-week ~ri~. S~ation of ~mples over the quarter yields the es~atefor a 13-week ~ri~. ~e par- ~c~ar 13-week peri~ is the one e~ending from • e 12~ week of the qua~er prec~ng ~e quarter ~ interviewing ~rough the 11~ week of the q~r- ter of inte~iewing, since mbMation is gear~ to weeks ~ inte~ewing w~ch He ~ ~e calendar qua~er. ~fle ~s ~ri~ d~s not corres~nd exactly with the 13 calendar weeks of ~e q~er, ~e ~splacement is s~ll, and estimate~ ~de • is ma~er are us~ as ee~ates for ~e calen~r quarter. Similarly pr~uc~ est~ates s~m~ for 4 successive qua~ers would yield an appro~- mate est~ate of hospital ~ys for ~e ~p~ation over the year. ~s estimate d~s n~ in~uds hos- pital ~ys for persons who ~ witch the 2-w~k ~ri~ immediately prec~g ~e week of inter- ~ew, since ~e eco~ of ~e hougehold sudsy • e Hv~g ~p~a~on ~ ~e week of lnte~iew. 18 A sec~r.d Lllu~raticn relates to c~r~ining timat~s for more tb~n I quarter, v,'h~n th~ q~.arter- ly es~mates l~.~ve ke~n e~res~o_d as rates. The lYroblem might he formulated in man)" ~ys. Cr.e ~I1 surf.ice h~re. From each quarter's ~-~mpl~ an estin-,ate of the average number of persons ~ho have exl~erienced 1 or more days of heal-disability in a 2-week period can he produced. This figure divided by the average popular.ion for ~e quarter yields a rate. An annual rate based on e.x-~rience for a year rather than for a quarter could be. formed in more than one way. An acceptable solution is a weighted average rate calculated as indicated: Let Bi be the number of persons in ~ quarter with 1 or more days of bed-disability in a 2-week period, as estimated in the first example above, Ni be average population in ith quarter, and hi equal to Bi/Ni be the quarterly rate; then the annual rate, R, may be estimated as 4 Z RiNi • R equal to 4 '$amplin.g and Measurement Errors Reliability of statistical surveys,--All statis- tical surveys, whether based on samples or at- tempted complete enumerations, are subject to potential inaccuracies. These risks include, among others, errors in conceptual formulation, ambigu- ities in definition and in the questionnaire, faulty classification, interviewer variability and bias, respondent bias and variability, biases from non- response or Incomplete coverage, mistakes in editing, and tabulation errors. This broad group of imperfections can be subsumed by the term "measurement error," which includes all nonsam- pling hazards. Measurement error plus sampling error may be called total survey error. Ideally it is desirable to detect all major components of total survey error, quantify each of them, and allocate resources in such a fashion that total survey error is minimized. Occasionally it is preferable to exclude from consideration cer- tain specified components, even if they are large, TI06653832

If k is, 1 2 the state-eat is true appr ox ~imate 2 times out of 3 19 tizaes out of 20 99 tiras out of I00 Reports published by the health survey include statements of sampling reliability for principal es- timates included in the report. In addition, as ex- perience is gained, it is expected that general guides and rules of thumb will be developed where- by users of the statistics can secure approximate sampling errors for other figures, with a minimum of effort. It may be useful to note relative magnitudes among some of the different classes of statistics which will come from the household survey. If V is the relative standard error for a sta- tistic which refers to an estimate for a U. So total, then relative standard errors for the same statis- tic when it refers to other subdivisions of the United States usually will he of the general mag- nitude indicated in table Bo Table B. Magnitudes of statistics for sev- eral types of area Area U. S. total A geographic section (e. g., New England) Rural United States The non-metropolitan urban sector of the United States Metropolitan United States Rough magnitude of relative sampling standard error V 3.3V 2.0 V 2.0 V 1.5 V Similarly, ff A is a relar.ive standard error for a statistic which rests on data for a year's in- terviewing, the magnitude of the corresponding relative error for the statistic based on I quarter's sample wi]/be about 1.7 A. If B__ is a relative error for a characteristic possessed by 1 percent of the population, the rela- tive error for a statistic possessed by I0 percent of the population will have magnitude approximately 19 T106653833

20 geree~t cf_~; fi:e relattve error for a s~fi~¢ ~tcd~ of ~ ord~ of tO ~rcent c,f B~ Smn~d e~ors of ~erence~ k~w;a~ e~- w~ b~ 4D F~rcent larger fl~an th~ ~nflard error of ~e sm~s~c at a fixed p~int ~ ~me. F~lly, ~e reliableS, ef an es~ated rate or Fercen~, computed by us~g sample ~ta for b~ n~erar~r ~d denominator, de~nds upon ~e siz~ of the rate and ~ size of the total upon w~ch the rateis based. Es~ared rates are rela~vely more reliable than correspon~n~ absolute es~mates of the numerator of the rate, part~cularl¥ if the rate is hi.~1. However, rafic~ of e~timated aggregates to to~./~;ula~cn for an age-~ex-co[or class have the &ame rela~ve sampling variance as the mated aggregate, as a result ef the ratio e~_n-nat- ing tec~mique ,~°hich was employed. Illustrative sarnplin~ errors.--Relative pling errors have keen calculated for a number of estimated national statistics based on data for the first 13-~'eeks of intervievdng. The extent to • ~'hich these values prove to he D, pical must a~vait the e%ffdence of later data. Illustrative errors are presented in table C. Table C. Illustrative relative sampling errors for national statistics from the U. S. Na- tlonal Health Survey, based on data from interviewing during the 13-week period ending September 29, 1957 Statistic Number of bed-days for medically attended chronic conditions in last 12 months ................................................. Number of visits to the doctor ..................................... Number of acute conditions ......................................... Number of acute cbndltlo~s, medically attended ..................... Number of persons with chronic limitation of activity .............. Number of persons injured in accidents ............................. Number of persons injured in motor-vehlcle accidents ............... Size of statistic (000,000) 756 199 70 47 17 14 i Relative standard error 0.010 0.022 0.030 0.042 0.030 0.051 0.175 ZO T106653834

NATIOHAL ttEALTH SURVEY B "2t T106653835

22 I"ABL~ & (A©cideaCs and |nJerles) Ill I I I I ..... TI08653836

(h) (11 o) (k) |1) ; I TI066538,37

Csrd A HATIOHAL HEALTH SURVEY ~heck List of Chronic Condition= t, ~sthr~ 16. K|dney stono~ or other Z. Any allergy kldeny trouble 3. Tubnrnulo~iS 17, Arthr]tls or rheu~tls~ 5. Repeated attacks of sinus trouble 19, O|obetes 11, Trouble wlth varicose voles trouble 12. E~errholds er piles 2~. Repeated trouble w;t~ 1~. Gallb1~dder or liver trouble b~ck or spine Card B RATIONAL HEALTH SURVEY Check Lis~ of lepairments Card C NATIONAL hEALTh 3URVEY For: ~orkcrs and other persons except housewives and Children 1. Cannot work at 811 at present, 2. C~n work but limited in amount' or kind of work, ~. Can work but lim|ted In kind or at~ount of outside activities. ~. Not l~ited in any of these ways. Card P NATIONAL HEALTH SURVEy For: Housewife Card E NATIONAL HEALTH 6UhVEY For: Children fr~ 6 to 16 years old and others goln~ to school Cannot gO to C¢nool st all at present tl~¢, Con 9o to sohnbl but limited to certain types of schools or in school attendance. Can 9o to school but li,lted in other ~¢tlvltle$. ~. ~ot l|m|ted in any of these way3. Card F ~ATIONAL hEALTH For: Chi]dren under 6 years old Card C HATIOKAL HEALXE 1, Confined to thO h~u;~ nll the t|me, except In e~roen~ie~. 2, Can 9o outslde but need the holp or another person In getting around outside. 3. Can go outside alon~ but have trouble In 9ottle~ around freely. ~. ~ot limited in any or th~e w~y~. Card H NATIONAL HEALTH ~UHVEY Fs~lly Income durln~ pssl 12 ~onths clu~ foot. 7. Cerehr~l palsy. O. Paralysis of any klr~. t. Cannot keep house at all at present. 2. Can keep house but limlted in a~ount or kind of housework. ~o Can keep house but llmited ~n o~ts|de aerie|ties. ~. Not limited ;n any of these ways. 1. Cannot take part at ali in ordlnar play with other children. 2. Can play with other ehi)dron but li~ited In a~ount or kind of play. ~. ~ot Tim[ted in any of these ways. 1. under $500 (Ir~lu~ln~ Io~) 2. $500 - $999 3, $1.000 - $1.999 $2,000 - $2,999 5. ~,000 - $~999 0o $7~000 - $9,999 O~ ~o

APPENDIX II ESTIMATING EQUATIONS In the National Health Household-lnterview Survey, the following algebraic statements sum- marize the estimation process for X' an estimate of X, a population characteristic. Let Phij be the probability of selecting the .th ith j PSU in the stratum in the hth Tab Area, with Phij --Ahij ' Ahi where Ahij is 1950 population of the hijth PSUand Ahi = 1950 popu- lation of the,hith stratum. and let Achij be number of persons in the cth color-residence group in the hijth PSU according to the 1950 Census. Then Lh 1 ch i=i j=l where Lh is number of nonself- representing strata in the hth Tab Area, V~h is an estimate of the numberofpersons inthe cth color- residence group in the nonself- representing strata in the hth Tab Area. The quantity Vch is the corresponding 1950 Census count. If next, ~vach is the sample aggregate of the X-measure for the ath age-sex-color class in the cth color-residence group in the nonself-rep- resenting strata in the hth Tab Area, and "~uach is the corresponding aggregate for self-representing hth Tab Area, and if further fh is the strata in the overoall sampling fraction for the hth Tab Area. then ~uach + ~vach Vchq is a first-stage ratio estimate of the characteris- tic for the ath age-sex-color class in the hth Tab Area. In precisely the same manner, an estimate of Zah, the current population of the ath age-sex- color class in the hth Tab Area is calculated as The total first-stage ratio estimate for the ath age-sex-color class is for the X-measure: X" E X" a = ah' and h for population: Z~ ; "ah h The final second-stage ratio estimate of the total X-measure is X' = ~.. X'~ Za, where Z ~" a a is the independent current population estimate for the ath age-sex-color class. 25 T106653839

APPENDIX ill SAMPLING AND MEASUREMENT ERRORS .Sampling Error -- Basic Formulation One of the attractive features of probability sampling designs is their inherent quality which permits determination of tolerance limits within which lie findings from the survey. More specifi- cally, for such designs, it may be determined, with any specified degree of confidence, what the maxi- mum differences are between results from the sample and those which would be found in a com- plete enumeration conducted under identical con- ditions. In simple designs, determination of sampling variance most commonly is made in a series of 3 steps: (1) An exact sampling variance formula for the design is derived mathematically in terms of (unknown) population pakameters. (2) Sample data for Individual units are used to estimate the needed but otherwise unknown parameters. (3) The esti- mated parameters are substituted into the derived formula, and sampling variances evaluated. For more complex designs such as that of the health survey, the procedure just outlined is usually not feasible or efficient, even when required form- ulas have been derived. Different methods, out- lined in this section of Appendix III and described in somewhat greater detail in the next, are used in the health survey, The fundamental rationale of these methods is simple and applies to all probability designs. All observations of a characteristic x are distributed randomly into m groups of k observations each. Each group permits making an estimate of ap- proximately-~ th part of the population total, by a sample design which is essentially the same as the over-all design. Thus if X~ is the estimate from the gth group, ~' the mean of the m values X' and X' is the over-all estimate, then g' m X' =~ g=l Xr is S~,, = X~, and the sampling variance of m S~,2-~, where S~ is estimated 26 variance of the group estimates: g m-iL m - (~') This general scheme of estimation has been rec- ognized by a number of statisticians. For example, Deming4 speaks of it as the Tukey plan; Hansen, Hurwitz, and Madow2 and others describe it as the random group method. It is being used more w/de- ly as electronic computers make it more practi- cable. .Sampling Error Functions The picture just sketched needs to be further highlighted in two important respects. In the health survey, attention usually is centered on estimates of aggregates or on estimates of ratios of two es- timated aggregates. In either case, since simple estimated aggregates are obtained as ratios of an estimated statistic to estimated population, the ul- timate estimate is a ratio, say I~ , of two other estimates, say, Y' and X'. Under the heading "Es- timating Sampling Variances From Survey Data." beginning on page 27. a procedure for determining variance of a quantity X' or Y' is presented. An entirely analogous procedure yields the covariance of X' with Y'. Finally, tel-variance of the estimate 1~' is obtained from the equation; where the V-symbols represent relative variances and covariance of the subscript variables. Thus the procedure can give variance for any aggregate or ratio. In the health survey, thousands of different estimates are being made. Even with high-speed computers, the cost of calculating var- iances far each separate estimate would be pro- hibitive. Further, such a step would be undesirable in that it would yield estimated variances which, TI0~653840

at ~s ~o he ~consi~e~t am~g ~s~Ives. Fo~ ~se r~s~ns, ei~ o~ of~ co'~s~s ~ followed ~ S~;ey p~li~dcns. In one cf ~ese, variances are ~]c~a~ed for o~y a few key izems, and ~m r~d~ is a~ov;~ ~o i~er from these ~e ord~ of va~e for o~er item. In ~e second co~se, a ~oup of variables ha,ring cerm~ common characteristics--such, for e~ple, as ~g binomial va~a~es--hut ~¢r- ~g in absolute size, are ~ed ~o establish a fi~ c~e w~ch expresses a "law" of valance for variables of ~e class. ~e fitted cu~e usually es~a~e, v~ , its rela~v¢ sampling variance, and a and b ar~ consents of ~efk~ed cu~e. kis ~ese rea~ngs from ~he curve w~ch are used as ~s~ estimates of the variances. Estimating Sampling Variances From Survey Data For calculation ofvariapces from sample data, the universe is divided into th.e 4 sectors displayed intable I. The contribution of each sector to over- all variance is computed separately. For sectors I and lit the sampling ratio in the first stage of selection is unity. Accordingly, the between-PSU component of variance for these sec- tors is zero. Table i. Sectors for use in calculating variances for a calendar quarter Sector number I II III IV Sector name Number Number of of se$- PSU'S merits Self-represent- ing SMA's ....... 77 703 Nonself-repre- setting SMA's--- 13 65 Self-represent- ing urban and rural PSU's ..... 28 90 Nonself repre- senting urban and rural PSU's- 234 692 The general scheme of estimating within-PSU variance for these sectors "is the random group method previously mentioned. It will be illustrated for sector I, the self-representing Standard Met- ropolitan Areas. X'n.~ se=n-n~nt is d~e muir c~ sampling wir/-dn PSU's.and accor~y is made th~ ~sis ~ ~cu- ladcns of ~i~-PSU variance. ~e 703 ~ers~p of approy~ately 85 s~gments ~ch is dory dete~ned. ~e selectisn proces~ is con- pan of each of the D~s of accents in ~e sector. ~e n~rs of semen,s by D~s are: Total Central City ................ Urban fringe ................ Other urban places. Rural ....................... New construction areas ...... Number of segments 703 354 156 30 84 79 Inflated totals fora characteristic Y for each group are established, with the summarizing operations: k Y'g =~ Y'. i=l g~ where Y'. is the estimate for that part of the uni- gt verse which is represented by the ith segment in the gth group, and k is the number of segments in the gth group. The variances of Y'g and of Y' are calculated as are those for X' and X' respectively g on page 26, so that Irn m m p.] ~ 2 = ~-II Z (¥,.)2 - <~ sy, gol g--l J The contribution of sector III is calculated in the same manner. For the nonself-representing sectors, an ulti- mate cluster technique is employed in calculating variances. Further, since there is but 1 PSU in each stratum, the strata are grouped into pairs, placing silrdlar strata in the same pair. This proc- ess is described as a collapsed strata technique. It is illustrated for the nonself-representing urban and rural PSU's. Data for the 254 PSU's are consolidated into 117 pairs, Yi being the estimated totaI for that 27 T106653841

part of th-~ p~p_ulaticn represented hy the ~th Fair ar.:l Y'. l-2inl that F-~xt of fi~ tota/ e~imat~d by the ith PSU in the ~th pair. Variance of Y' timated as g' L g 2 where Pgi is 1950 pepulation of ith PSU, Pg is 1950 population of gth group and L is the number g of PSU's in each group. Since all groups contain 2 PSU's, Lg is a constant equal to 2. Further, since , = y' y' Yg gl+ g2, the sampling variance of 2 2 Yg Sy'gi' and the total for the sector, Y' , has variance 117 Sy, ; Sy, g~l g The ~ame prccedure is uz~.A for sector II. Variance of rite stu-vey total is simply the s~.wn oi[ th~ varian~e~ for the 4 sector~. Measurement Error Measurement error can be divided into com- ponents in a varieD" of ways. One useful scheme is to separate it into bia.__~s and nonsampling variance. Nonsampling variance has in t-urn many compo- nents. Among these are variations which have their source in respondent, interviewer, classifier, edi- tor, or tabulator. The method of estimating sam- plin~ variance which is used in the health survey includes most of the measurement variations, al- though it does not include those components of verier.ion which are unaffected by the size of the sample. With some exceptions (found in edi[s for consistency~, the biases of measurement, from whatever source, are not treated in ~he present report. The mai, text of the report, on page 19, lists several routes being taken, all intended to im- prove evaluation of measurement error. An ulti- mate goal is establishment of a model for analyz- ing over-a]l error and its components, and for guidance toward efficient use of resources in min- imizing total error. Z8 Tl06653842

APPENDIX IV STRATIFICATION OF PRIMARY SAMPLING UNITS Principles A twin objective of many sampling designs is a pattern in which individual primary sampling units are as internally heterogeneous as possible, and in which each stratum formed from grouped PSU's is as homogeneous with respect to PSU's as possible. Said in another way this means that ulti- mate sampling units withina PSU should tend lobe unlike one another, but that PSU's within a stratum should tend to be alike one another. This twin ob- jective was sought in the Health Household-lnter- view Survey. Three broad specifications of the survey molded the main outlines of the modes of stratification in the NHS. These were: (1) The requirements of end product which were that ,separate estimates be pre- pared for major Standard. Metropolitan Areas, for a number of geographic sectors of the country, and for differing densities of population (metropolitan, other urban, and rural). (2) For administrative reasons, and inorder to minimize operating costs, stratification in the NHS was to be coincident, inso- far as feasible, with that of the Census Bureau's Current Population Survey. (3) The general charac- terization of the stratifying process was that it produce relatively homogeneous socioeconomic classes of PSU's--with rhis term being further interpreted to reflect geographic location, density of population, rate of population increase between 1940 and 1950, proportion of nonwhite, type of in- dustry in predominantly urban areas, and type of farming in the rural areas. Within these specifications, the approzimately 1,900 PSU's were classified into 372 strata, the following rules serving as principal further guides in the process. In each case the rule is presented as a positive statement, although obviously there hadto be some compromises among rules in order to produce a desirable result. 1. Except where a single PSU was larger than an average stratum--size being measured here as elsewhere in the stratification process by 1950 population--s~ata were of approximately the same size. This meant about 300,0~0 persons to a stratum. 2. Since the general design called for sample selection of a single PSU from each stratum with probability proportionate to size, each PSU with the population above a lower cutoff became, by it- self. a self-representing stratum. The effect of all rules was to set this cutoff at 400,000 (1950 pop- niation). 3. Also included as self-representing or cer- taint3, areas were any Standard Metropolitan Areas with the population somewhat less than the cutoff, but within 100 miles of an SMA above the cutoff, The rationale was that the same field organization which served the larger city could also serve the other, and thus reduce costs. 4. Solution of the allocation problem (page 31) led to the conclusion that a nonself-representing Tab Area--that is a Tab Area not made up entirely of self-representing PSU~s--should contain not less than 4 sample PSU's ff it were a Tab Area of Standard Metropolitan Areas, and not less than 8 sample PSU's otherwise. This meant in turn that such Tab Areas would contain corresponding minimum numbers of strata and this fact influenced ultimately the number of different strata which were formed. 5. Since end-product specifications required, for purposes of comparative analysis, both urban and rural Tab Areas within each geographic section, it was decided to make the first stage of sample selection identical for the other urban and the rural Tab Area within the section. Thus, each PSU drawn from other than Standard Metropolitan Areas be- came the first-stage unit for I urban Tab Area and 1 rural Tab Area, and 2 sets of ultimate stage units or segments--1 for each Tab Area--were drawn from each such PSU. This step had to be taken into consideration later in calculating variances, since first-stage selection for these Tab Areas was not independent. 6. Stratification proceeded in a sequential man- ner: tentative classification with respect to 1 major specification or rule being followed by ten- tative subelassification by a second rule and then by further subclassification by a third. As the proc- ess continued, occasional changes in the first ten- tative classifications had to be made. After semi- final stratification was completed, there was a re- view of results, anda few subjective changes made which reviewers thought would increase socio- economic homogeueity betnveen PSU~s within strata. This intreduction of judgment in the stratifying phase of the survey could, of course, produce no bias. If it was well done, it reduced sampling vari- ance; if it was poorly done, at worst it would in- crease variance. 29 T!06653843

ru~es led to a classi~cadon of ~h~ approxhnately 1.900 PSU's into572 s~ra~a. Of[hess. 110 are com- posed of a single s~-representing PSU. Col|ec- ~ively, these II0 strata represent 52 percent of [he population in ~he universe. For [hem there is ns hesveen-PSU component of variance. The re- mairdng 262 strala vary a great deal among one another, some b~ing metropolitan, some urban. some rural, and all obviously exhibiting sti|l other differing features as a consequence of ~he stratification. Even so, 3 examples of actual s~ra~a formed may contribute to a "feel" for the nature of nonself-representing s~rala in ~he health survey. Example A. Sparsely populated stratum PSU ' s (defined by counties) Total ............... Coconino, Ariz.---/ ....... Dons Aria, N. Mex.*--u Graham, Ariz.- ............ San Juan, N. Mex.- ........ Valencia, N. Mex.- ........ Nava~ o, Ariz.- ............ Uintah, Utah .............. Alamosa-Costilla, Colo.--- Mineral-RioGrande, Colo.- .................. Montezuma, Colo.- ......... Montrose, Colo,- .......... Pinal, Ariz.- ............. Preliminary 1950 population 254,235 23,755 39,044 13,018 18,116 22,574 29,263 10,259 16,572 13,330 9,937 15,024 43,343 *In e~ch of the three e~amples, the starred PSU represents the stratum in the sample. Ex~_---ple B. Moderately densely populated non-M=_~r opolitan stra~ PSU's (defined by counties) Total ............... Harrison-Heard-Troup, Florence-Marion, S. C.---- Baldwin-Jones-Twigg, Ga.-- Calendar-Sumter, S. C.---- Prelir~nary 1950 population 315~623 68,008 112,208 45,580 89,827 Example C. A nonself-representing SMA stratum Standard Metropolitan Area Total ............... Springfield, Mo.- ......... Sioux City, Iowa .......... St. Joseph, Mo." .......... Preliminary 1950 population 301,[06 103,959 93,629 5O T106653844

t APPENDIX V, THE SAMPLING ALLOCATION PROBLEM Leading Considerations, A fundamental fact which conditions the de- sign of a multipurpose survey and the allocation of resources is that no single factor will determine uniquely the design, but rather a balance must be sought taking into consideration leading objectives. In planning the Health Household-Interview Sur- vey, leading considerations were identified as fol- lows. 1. The survey was expected to provide sepa- rate estimates for anumberof geographic sections and for metropolitan, urban, and rural sectors. This condition was converted initially to a provision that separate worksheet estimates be produced for each of the defined 41 Tab Areas, although the Tab Areas would be consolidated into a lesser number of groups for most purposes.'. 2. A household survey was predicated, which in the United States ordinarily means a multistage area design. 3o Tentative determination had been reached as to target sampling tolerances for estimates which were to come from the survey. 4. Preliminary study of requirements and re- view of probable administrative and operating costs strongly suggested that initially the structure of the health survey should parallel in large measure the Current Population Survey (CPS)which was also a general-purpose survey of households. Sig- nificant savings might be possible, if the 2 surveys were companion undertakings. 5. The survey was to be a continuing activity, geared to production at quarterly intervals of na- tional estimates of characteristics of high incidence, and production of other statistics for the Nation and for parts of the Nation at annual intervals. 6. Appropriations set budget limitations on the design. Outline of Design Solution The specifications suggested that equal relia- bility be sought for estimates for each Tab Area, The target tolerances and previous design experi- once suggested further that a multistage survey could be designed which would meet requirements and which would contain a possible 700 to 1,200 households per year per Tab Area. Experience with CPS indicated that a total of 300 or more strata with 1 sample PSU in each stratum were desirable. Since the principle had been adopted that the 2 surveys were to be com- panion activities, and since the CPS was operating with 330 strata, it was decided as a first step to adopt tentatively the CPS stratification for the NHS. This tentative decision was reviewed and modified in a later step. The budget factorwas now introduced. For the tentative design, which was beginning to shape up, it seemed that about 36,000 households per year, or a little under 900 per year per Tab Area, was feasible. At this point, the precision requirements for each tabulation area were considered in terms of the components of variance. The set of strata for CPS in each tabulation area was examined to see if they were adequate to meet precision require- ments for the Tab Areas. In the areas in which the minimum stratum requirements did not appear to be satisfied, additional strata were created, thus bringing more PSU's into the sample. In some cases this was accomplished by splitting an ex- isting stratum into 2 parts, letting the PSU which is in the Current Population Survey represent the part of the stratum in which it falls and selecting a new PSU in the other part. In other cases, it was necessary to rearrange some strata to pre- vent great variation in strata sizes or in the urban- rural composition of a stratum. In such cases new PSU's were selected, and as a result 69 of the PSU's for the CPS are not included in the NHS. An additional 111 PSU's not in the CPS were selected for the NHS sample, A principal tool utilized in carrying out the anal- ysis indicated in the previous paragraph is ex- pressed in the approximate relationship x 2 2 , where V , = VB + VW m n V~ is between-PSU tel-variance in the population, V~2v is within-PSU tel-variance in the population, Vx2,, is sampling tel-variance of as estimated char- acteristic m is the number of PSU's in the sample for a Tab Area. and TI06653845

n is ~_~ nurc~n~r ~f hc--s~eho!c[~ in the sample for a Tab Area Value~ of V~and V~g were calculated for anuml:er of household statistics from r.he CPS and other sm-veys. 'v~,,. and n were set from first appropria- tions .~et by joint consideration of target tolerances and budget, For each of the eeveral hotmehold sta- tistics a valu~ of m was calculated, using the above equation, for nonself-representing strata. Using "t~'pical" solutions, this step determined the needed num]:er of PSU's in each Tab ~rea a~fl consequ.=~ntiy the numl:_or of sxrata ~'hich ~hculd h=- establish~-d, since I PSU was to ha drawn from each stratum, ~esult The consequence of these actions is the health survey sample design, which was planned to have 372 strata, 372 PSU's, 41 Tab Areas, and 36,000 households with 115,000 persons in it each year. As noted elsewhere in the report, the original allocation of resources will ~e modified as con- sumer interest and experience dictate. 52 T[06653846

APPENDIX Vl ILLUSTRATION OF DRAWING PSU'S AND HOUSEHOLDS INTO THE SAMPLE Selection of Primary Sampling Units Section 5 of this report outlines the main fea- tures of sample selection in the health survey. This Appendixlllustrates the principal steps of that process. Assume a particular stratum contains 4 pri- mary sampling units, or PSU's. These are listed, together with their 1950 population, and cumulated population, as in table 2. Table 2. Primary sampling units in stratum number 428 PSU Cedar Rapids, Iowa, SMA ................... Lincoln, Nebr., SMA .... Topeka, Kans., SMA ..... Waterloo, Iowa, SMA .... 1950 popula- tionI 104,000 118,000 104,000 99,000 Cumula- tive 1950 popula- tion 104,000 222,000 326,000 425,000 1preliminary sod epproxlmste population £igures are used in this example. A random number between I and 425,000 is select- ed. Assume the number is 301,265. This number selects Topeka, Kans., as the sample PSU from stratum 428.* *In three respects, the example is a streamlined version of detailed seleetion.~(1) Where the stratum in the health survey and in the Current Population Survey were identlcal, thePSU drawn earlier for the CPS w=s used also in NHS.- (2) Where a CPS strotum was divided into 2 strato in NHS. an unbiased selec- tion procedure retained the CPS PSU for one of the new strmtn. (3) Those PSU's which are found also in CPS were selected initially with probability pro= portional to size, and also under restrictions of the Goodmsn-Kisb controlled seleetion technique which increases the probabilities of selection for preferred combinations of units." Selection of Enumeration Districts and Segments The exact procedure for selecting segments varies depending on whether the Tab Area involved is a Standard Metropolitan Area, an "other urban" area, or a rural area, but the nature of the procedure is the same for all areas. It will he described for a typical metropolitan Tab Area for which not all first-stage sampling units were self-representing; i.e., for a Tab Area in which there is more than 1 PSU in the sample. In following this selection proc- ess it is useful to remember that the final sample of households and persons is intended to be self- weighting within the Tab Area, which means that every household in the Tab Area has an equal chance of being selected. Assume that this Tab Area has 5 PSU's in the sample, 3 of which are self-representing, and 2 of which are not. Since the over-all design has an avesage annual sampling rate of about 1 in 1,400 and since 144 segments are to be selected from the Tab Area, assume this typical Tab Area contains an estimated 200,000 segments in the population (page 13). More precisely, the assumption is that the Tab Area contains 200,000 size ,.rneasures, where a size measure is equal to 6 households, and the number of size measures is the number of households in 1950 divided by 5. The first step is to allocate the 144 sample segments to the 5 sample PSU's. This is done in proportion to the estimated size of the stratum represented by the PSU. For example, if a partic- ular sample PSU contains 5,000 size measures, and was drawn from a stratum containing 25,000 size measures, it represents .those 25,000 size measures in the sample and, therefore, represents one-eighth part [25,000 divided by 200, 000] of the population in the Tab Area. Therefore 1/8 bf 144, or 18 segments are assigned to that PSU. In order to facilitate continuous sampling, and to reduce costs by having samples in adjacent quarters also geographically neighboring, 4 quarterly samples are drawn simultaneously, as sketched in the next paragraph. Accordingly, the 18 segments are di- vided among the 4 quarters, so that either 4 or 5 segments will appear in each quarter. TI06653847

~.~ n~xt ~tep i~ to localize ~e ~arnple ~o ~ ~er ~n ~ P~U. For ~ 19~0 C~u~, vary ~dy ~ s~e, ~ut u~y c~n- Sin n~ less ~n 10 ormore ~ I~ size meas- ures. Ass~e ~ ~usIradon ~[ ~e selec[ed PSU ~ ~,COD s~e m~s~es conmi~ ~0 ED's. For ~e PSU cf ~e ~mple~ ~ ~e~ts ~ requ~ ~ some q~ers and 5 ~ larger of ~ese n~hers is idendfi~ as ~e num- ber of "sm~g point." ~us in t~s P~ ~ere are 5 sm~n~ points. It is intend~ ~at ~ese smr~n~ points ~ ~s~u~ed rando~y, bu~ sys- ~ema~y ~ough~ ~e PSU and ~at ~ey sere as selectors ~ ED's and semen,s for the f~rs~ quarter. ~is ~s done in the follo%~n~ manner. ED's are arranged in sys~ema~c sequenc~ cen~r~ c1~y ED's ~s~f~rs~, followed by all urban- ~zed fringe ED's, and ~hen by o~er urban ED's, and finally by r~ ED's. ~efirs~ sta~ng ~int ~s determined by ch~s- in~ a random n~r ~tween 1 and i;~0 s~ze measures ~n ~he PSU ~v~d~ by 5, ~he num~r of s~ardng po~n~s]. Say ~s n~r is 725. ~en • a~ Hs~ed ED w~ch contains ~e 72S~ cum~a~d size measure ~s ~nclud~in ~e sample, as are also ED's ~th ~e 1,72St~, 2,72S~. S,72Sth, and 4,72S~ c~at~ size measles. Consider the ED ~ ~e 72S~h s~ze measure. Suppose it contained 100 size m~sures, ~den~ied ~ ~he c~a~ed listing as nu~rs 70S ~rough 804. ~ pr~ess jus~ desczi~ l~tes ~e firs~ s~ardng point ~en not o~y in th~s pardc~ar ED, but at the 21st size measur~ [random n~r 72S, minus 70~, plus I]. Ma~ng use of Sanborn" and other de~a~ed ~ps, ~ ED ~hen ~s "accented" on a new map into i~ ~ts appro~a~ely eq~l in s~ze (i.e., in the n~r of e~ec~ed households). ~ese units are n~bered consecuUvely from I ~rough I00 in a systematic fas~on ~in~n~ w~th a rando~y lo- cat~ s~arL ~e ~ or segmen~ num~r~ 21, con~a~ng an e~c~ 6 hous~olds, ~comes sample segmen~ for ~e firs~ qua~r of ~n~er- v~ewin~. ~s same pr~edur~ ~s ~rried ou~ for o~her chosen ED's in ~e PSU and for o~er sample PSU's in the Tab Ar~. It w~ ~ nodc~ ~at, ~use some n~rs are nor e~c~ly d~sible by o~ers, in the e~mple PSU Z rather than ~e calculated 4.~ s~ents are in~e~ew~ ~ the f~s~ qua~er. M~morandum rec- ords are maintained so ~ ~er the Tab Ar~ e~cdy i/4 of 144 or 36 segments are in~e~iewed each quarter. "Published by the Sanborn Map Co,~ New York. N.'Y. 34 In ~-~ ~ample ED, the 21st se~n~nt ~'as inter- viewed ~h_~ ~ar queer. For ~e ~eccz~ q~ner, + ~ j = 97~ ~z~ent, is ~ ~ sample; ~e four~ q~er; ex~ep~ ~t ~ memor~d~ record aga~ is ~ed to ass~e ~ ~y 18 seg- men~ from the PSU are ~clud~ over the y~r. In ~e follo~ng year, se~en~ are select~ ~ such a manner ~at ~ey are geograp~cally hair.ring the secants in ~he ~rs~ sample at a~ut the same time of the year, Thus i~ is tha~ ~¢er ~e y~r, for ~e s~a~m fromwhich ~e ~mple PSUcomes, ~e probabiliw • a~ any segment, household, or person is ~ the sample is ~e pr~uc~ of the prob~i~ of selec~g • is pmrdc~ar PSU (5,000 ~ded by 25,~) times • e probabili~ of selecting a ~rtic~ar segmen~ wi~in ~e PSU (18 ~vided by 5,~); or m other words is I/5 ~mes 18/5,0~, w~ch is 0.~72. By virtue of the way in which ~ sampl~ was dis- tribute, this is exac~y the designed over-all sampling portion, 144/200,000, for ~e T~ Area. ~e pro~blli~ for any ~rson from the e~mple T~ Ar~ appearing in a Wen quarter is approx- imately 0.~018. Variations of Detail The principles .of selection were uniform throughout the survey. Dependin.g upon the partic- ular areas which fell .into the sample and upon the types of resources available for those areas, additional steps sometimes were taken in the lection process. For example, detailed block sta- tistics were available for many cities. In these cases, a selection of blocks proport/onal to size was made within sample ED's before making a direct selection of segments. In some instances a block was further subdivided and subsampled be- fore final selection of segments. If it was found from a Sanborn map or other source that the pro- spective uhimate sampling unit was a large apart- ment building, still another stage of subsampling was introduced to bring the final unit closer to an expected 6 househ.olds, In some cases, the selection of samples in Washington results in the inclusion of a segment in which the field lister or interviewer finds many more than 6 households. This may occur because of new construction unknown in Washington, or be- cause sampling materials were incomplete or in- accurate. In instances in which the segment ob- viouslyappears tocontainmorethan 20 households, field manuals give detailed instructions for sub- sampling the segment and interviewing only the subsampIe, in a manner which reduced costs but avoids introduction of bias. A price of siighfly high- er variance is paid whenever this be'comes neces- TI08653848

APPENDIX VII RANDOMIZING ASSIGNMENTS, AREAS, AND WEEKS Basic samples in the health survey are drawn to represent the population of the United States over a calendar quarter. It is efficient in terms of operating procedures and reduction of variance, and furthermore, desirable in terms of potentially available end product, to make each week's col- lection a random sample of the population. This is done. The randomization of assignments, areas, and weeks is quite an elaborate process. To follow the process through in all its detail most readers would find tedious. For this reason, a description is given by means of an example which exhibits leading features of the process while omitting a number of lesser details. Dimensions of the Problem, For administrative reasons (which inthe main are consistent with minimum costs) a given inter- viewer operates within a single geographic section --with a few exceptions--and usually within from I to4 contiguous PSU's. Consequently, randomiza- Uon of assignments, areas, and weeks was carried out separately within each geographic section. In this process the 8 largest SMA's were ex- cluded from the sections and treated separately. There are 11 sections in the country, each divided into 3 tab areas: metropolitan, other urban, and rural. Each Tab Area contains 36 segments for the sample for a quarter, and thus a section has 108 segme.nts each quarter. There are a total of 120 interviewers to cover a gr.and U. S. total of 1,476 segments per quarter (including the 8 largest SMA's), Thus, on the average, 1 interviewer covers 12 segments per quarter, Excluding the 8 largest SMA's, the 108 segments per quarter in a section require an average of 9 interviewers for the section. A typical assignment for an interviewer for a week is 2 segments or an expected 12 households to he interviewed, although an assignment may consist of either 1 or 3 segments. An interviewer may or may not have an assignment in a given week. She nev- er has more than 1 assignment in a week° Thus, the typical situation in a section over a quarter encom- passes $4 assignments, 3 tab areas, and IS weeks, with 6 assignments per interviewer, although the assignments per interviewer may range from 3 to 13. An effort is made to provide at least 1 assign- ment to each interviewer each month, in order to avoid having too great a time lapse between inter- viewing experiences. The objectives of intraquarter arrangements are: I. Obtaining approximately equal representa- tion from each of the 3 Tab Areas in each section in each week 2. Spacing thework of each interviewer at ap- proximately even intervals over the quarter, and 3. Randomizing assignments (segments to be interviewed) over the weeks of the quarter. Principal features of the way in which these objectives are reached are illustrated in the fol- lowing numerical example of a composite geograph- ic section. It should be observed that there is no unique way of accomplishing the objectives and that the method chosen is but one of several possible methods. Example This geographic section contains the usual 3 Tab Areas: metropolitan, urban, and rural, each of which has 36 segments to be interviewed over the quarter. Nine interviewers have been hired for work in the Census Region which contains the section. The Census Regional Offices, of which there are 17, have indicated for each of the interviewers in which of the 20 PSU's in the sample in the section they can serve. This information has been reported to Washington (table 3). Table 3. Interviewer service areas Inter- viewer Can serve in PSU(s) numbered i, 2 3 5, 6, 7 8, 9 i0, ii, 12 13, 14, 15 16, 17, 18 19, 20 TI06653849

Format/on cf assi~Ernents in each PSU.--Th~ into a~si~men~ of 2 s~n~s ~(w~ i as~i~- meat conm~z eider 1 or 3 ae~ents number of ~oupin~ is to puz ~ike s~ents in si~men~ and to oh:sin ~Isnce he?~'een ~n ~d rur~ Tab ~eas is i~us~a~ in tans 4 for a non-SMA prima~ sampl~g u~t which contains 6 segments; the seg- ments co.acted by a line ~ing metro of the s~e assig~ent. Table 4. Formation of assignments in PSU number 5 Segment number Urbanization classification ~5-I 5-6 5-2 5-9 u-5-14 u---5- Ii Urban segment Urban segment Rural segment Rural segment Rural segment Rural segment Thus 3 assignments are identified for this PSU. In PSU's that are SMA, the arrangement is in se- quence by central city-, urban fringe-, other urban places-, and rural-segments. Determination of number of assignments for each interviewer.--The number of assignments in each PSU having been determined, the number of assignments for each interviewer is established readily by zeference to the field report reflected in table 3. A new worksheet, table 5,is setup com- bining these two pieces of information. The first figure in each cell is the identification number of the PSU and the second figure is the number of as- signments in that PSU. The columns headed total number of assignments, SMA, and non-SMA are utilized later in the allocation process. Spacing interviewer assignments throughout the quarter,--The next step is to distribute the sum- her of assignments by week throughout the quarter in such a fashion that each interviewer's work is spaced at approximately even intervals over the quarter and so that the total number of assignments is rougifly constant from week to week. This step is carried ou~on another worksheet shovm in table 6. 36 Table 5. N~r-_her of assigr~en=s for ea~ inte~¢iewer Inter - viewer A 1-2 B 3-3 C1 4-1 D 5-3 E 8-3 F 10-4 G 13-4 H 16-2 J 19-2 Eumber of as- signments by PSU 2-3 6-3 7-2 9-4 11-2 12-3 14-3 15-3 17-3 18-2 20-2 Total nu--_h er of assigr=ents Non- All SMA 5 2 3 ~ 3 i I 8 3 7 7 9 3 6 i0 3 7 7 2 5 4 2 2 lIntervlewer C has in this example only I assign- ment in the quarter for this section. She has addi- tional assignments in other PSU's in a neighborlng section which were assigned because the locations were more accessible to her than tolnterviewers from the other section. Table 6. Spacing interviewer assignments by week Week number I 2 3 4 5 6 7 8 9 i0 ii 12 13 G G G F G G G F G G G F G F F D D F F D X F F D X D D X Y X D X Y Y D X Y A X Y A J A Y A J A J Y J B C B B The interviewer with the largest number of assignments--Interviewer G with 10 assignments in this section--has her assignments located by week on the first line of the table. Since she has work in 10 of the 13 weeks, she has an assignment in each week except for 3 evenly spaced and ran- domly chosen weeks. Note that at this point the ~ of each assignment has not been deter- mined, but only the fact that Interviewer G has an assignment in the specified week, Then the intemriewer ~:'ith the next largest number of assignments--interviewer F with 9 as- T106653850

si~ummuts--has her v;ee]~ of ~ork p~ed to t~ble 6. ~s is do~ hy en~e~g h~r id~n~i~den in 9 of ~e rem~g ~ cells of ~ ruble, m~ng and line and s~ attemplug equ~ spacing of ~e 9 assi~men~s. ~ ~a~cul~r, F is nol allo%ved ~o Prove 2 assi~men~s in ~e same week, ~is pr~ess is con~nu~ for ~ch inte~iewer ~I ~e 54 as- si~enZs for the section h~ve ~ place. ~e assi~en~ of In[erviewer C ~o week 13 was ~de wi~ consideration heinggiven also lodming of her assignments in ~e neighbor~g section. Interviewers E and H ~ch ~ve 7 assi~ent& In ruble 6 desi~a~ons X and Y have ~en us~ ~ lieu of E and H wi~hou~ decision as to w~ch is which. This decision is rese~ed ~o a later point in order ~o permit grater fle~bHiW in.placing work. Randomizing @ssig~en~s.--~e remaining problem is to match specific assi~ments randomly with wee~y allocations of wor~d for the in~er- viewers. An impor[ant side condkion is imposed on this process. As nearly as possible each week's sample is kept balanced by S~ assi~ments and non-S~A assignments. In ~his e~mple, wi~h ~4 assignments to be made during the quarter, either 1 or 2 SMA assi~ments will ~ made ~ch week, ekher 2 or $ non-SMA assi~ments, and a ~otal of 4 or 5 as- signments each week. Assi~ments first are made ~en~a~ively, and in a few instances it may become necessary for an assi~ent which has ~en allocated to one inter- viewer ~o ~ reassigned later in ~he process to another inte~iewer as ~e sequential assi~ment process reduces degrees of freedom in all~adng wor~oads. Before beginning the randomization of assignments one needs to assemble the da~ from ~ables 5 and 6 and from a new ~aNe--[sble 7. Table 7. Designation of PSU's and assign- ments as SMA and non-SMA ! Assignments in [ Assignments in these PSU's I these PSU's are SMA segmentsare non-SMA segments i, 3, 6, 12, 2, 4, 5, 7, 8, 9, i0, 15, 16, 19 ii, 13, 14, 17, 18, 20 Assignments within a PSU are then identified by a letter prefixed by a PSU number; e.g., the 3 assigmrn~nts in PSU Nurr~ 5 are 5a, 5b, and Table 8 r~ects nhe ~ ~c~on an~ ~n- d~m~a~en of as~i~en~. ~ne de~i~on ~ ~ cell ind~es fi~e in~e~e%~'er ~fl fl%~ s~ic as- silent ZO her in d%at week. Prc-c~d~e for ~ in ~ mbleis outed in~ re~ ~ara~ap~ of ~s Ap~end£x. ~e ~nidal deteminad~ is n~er of Sk£~ as- ailments f~ ~ch week. As noted earlier, ~s mus~ ~ eider 1 or 2 for ~ch week. %~ch w~ks ~et 2 is de~ermin~ randomly, excep~ ~h~ wee~ 4 and 8. which are to have a ~ota] of 5 assi~ents, are ~ve~ 2 SMA assi~en~s each, ~is action de~er- m~es also ~e n~r of no~-Sk~ assi~en~s for each week and ~ese are posted ~o ~able 8. All~ado~ is made firs~ then, for ~e S~A as- si~men~ forweek i. Table 6 shows~at interview- ers G, F, D, and Y are scheduled to work in ~he first week and ~e sin~e SMA allotment could be given to any one of ~e S~ assi~men~s ass~iated with these inte~iewers. Interviewer Y is no[ yet id~i~ as to wh~ther she is E or H. Collectively, G, F, D, E, and H accountfor ii SMA assignments. ~e of these is picked at random. ~e assi~ment picked was 6b, which also selects inte~iewer D. The entry D6b is posted in the first cell in week I. Two SMA assignments are req~red for the second week, to ~ given interviewers G, F, X, or A. ~e assi~men~s are nex~ selected rando~y from the SMA assi~ents available, as in week I. ~e assi~ents proved to ~ Alb and Fl2c. ~is process is continued for successive weeks. In drawing for week 6, ass1~ment 16b was selectS, and ~hus X was determined to be H, and Y to ~ E. It happened that when week 12 was reached only SMAassignments Ale, B3b, a,d B3c remained available. Since B could not handle 2 assi~ments in week 13, Ale was assign~ toweek 13, alongwfth B3b, which was drawn a~ random from B3b and B3c. ~e remaining assignment B3c, went to week 12. When the SMA assi~ments had ~en allo~ed, • e non-S~ alloca~ons were undertaken, ~gin- ~ng with week I, and using the s~e procedure as for SMA assi~ments. ~e drawings were such that in the eighth week a non-SMA assi~men~ wo~d have ~en allotted to in~ervlewer B. However, ~ere was none available to B who had be~ ~ven all her assi~ments ear- lier-she served only SMA ~erritory. Since she had se~ed inlieu of G, D, or H inweek 13 for SMA assignment, a random non-SMA ass1~ment from ~ong ~ose still available to G and D was s~- adjured for B in week 8. It turned ~t to ~ D7b. Two other similar changes had to be made to com- plete the panel. 37 TI06653851

Table 8. Final assi~T~ents Week nurser i 2 3 4 5 6 7 8 9 I0 II 12 13 D6b Fl2c Jl9b FI2b Glbc Hl6b D6a F12a D6c Fl2a Glba BSc Ala Gl3a Alb G14b B3a E8a Gl3d Jl9a Hl6a Gl3b Gl4c J20b Fl0a B3b F10e GI3c Dba A2a D7a Fl0d E9b EBb Fl0b HlTa Dbb Hl7c Fllb E9c Hl8a E8c Hl7b Flla A2e Gl4a A2b J20a E9a Hl8b E9d C4a Dbc DTb Total SMA i Assi Total Non- 3 Ass Total 4 All weeks 2 I 2 I i 2 2 i I i i 2 18 2 3 3 3 3 2 3 3 3 3 3 2 36 4 4 5 4 4 4 5 4 4 4 4 4 54 TI06653852

APPENDIX Viii SELECTED STATISTICS ABOUT THE SURVEY For ready reference, and for their, value in giving quick insight to various features of the health survey, there are assembled in t.his Appendix sev- eral tables of stads~cs on ~he survey (rubles 9-15). In most instances the figures which are shown are Table 9. Summary statistics on components o£ NHS Item Counties and independent cities ................. .- .... Primary sampllng units in population .................. Primary sampling units in sample ...................... Strata ........................ In national sample in 1 year: Persons ................... Households ................ Segments .................. Tab Areas ..................... Large SMA's which are sepa- rate Tab Areas .............. Geographic sections ........... 3,100 1,900 372 372 115,000 36,000 6,000 41 8 11 Table i0. Size of national sample for different time intervals Type of unit Persons ...... Households--- Sesments ..... PSU's ........ Number of units in year quarter 115,000 29,000 36,000 9,000 6,000 1,500 372 372 i week 2,200 700 115 about 60 rounded and approximate since they are in~ended to convey an impression rather than to serve any operational purpose. As a result detailed figures are not always consistent ~th totals. Table Ii. Size of sample over I year Number of units in sample over 1 year Type of unit Persons--- House- holds .... Segments-- PSU's ..... National total 115,000 36,000 6,000 372 Each geo- graphic section i0,500 * 3,3001 550* 34t Each Tab Area 2,800 880 145 (2) |AveraEe. 2Urban and rural Tab Areas In a given sample are represented by the same PSU, There is an average of about 18 different sample PSU's for each of the non-SMA, first-stage selections for Tab Areas." 'Table 12. Sector Approxima=e over-all sampling rates on an annual basis Approximate inflation factor (reciprocal of over- all sampling rate) U. S. total ......... New York SMA ........ Chicago SMA ......... Typical other large SMA ................ Tab Area with high- est sampling rate-- Tab Area with lowest sampling rate (NY)- 1,400 4,700 2,000 1,000 350 4,700 39 TI06653853

Table 13. Data on field ~upe~:isors and inte~-iewers Number of field super- visors ................... Number of interviewers .... Typical interviewer work- load in i week- .......... Typical interviewer work- load over 1 quarter ...... Typical number of inter- viewers in a geographic section .................. Typical time required for interview of a household, including travel and call backs (but exclusive of supplemental inquiries)-- 17 120 12 households 72 households 60 minutes Table 14. Summary operations report on in- terviewing for 6 months' activity Item Number of listings assigned for interview ...................... Number of listings demolished, vacant, or otherwise not eli- gible for interview (Types B and C exclusions) .............. Net number of listings eli- gible for interview ............ Nonlnterviews ......... ~J ......... Percent of listings eligible ..... Percent refusal ................. Percent other (not at home,etc.) Number of households with com- pleted interviews .............. Number of persons in households with completed interviews ...... Number or per- centI 24,032 3,251 20,781 1,271 6.1 1.2 19,510. 62,046 llncludes approximately 7.5 percent more house- holds than were desiEned for the basic survey; ex- tra households used in preparing estimates for one part o§ the country,' 4O l[u--_her of ~SU's Self- Nonself- Geographic area Tota~ repre- repre- sent- sent- ing1 ing Total ...... 372 II0 262 Boston SMA ....... I i - New York SMA ..... 1 1 Philadelphia SMA- 1 1 Pittsburgh SMA--- 1 1 Detroit SMA ...... 1 1 Chicago SMA ...... 1 1 - Los Angeles SMA-- 1 1 San Francisco SMA 1 1 Other SMA's Northeast Region 28 21 7 North Central Region ......... 37 27 i0 South Region .... 46 36 10 West Region ..... 15 14 I Other non-SMA PSU's Northeast Region 34 9 25 North Central Region ......... 70 i 69 South Region .... I07 i 106 West Region ..... 36 2 34 1In detail 9 se[f-representin~ PSU's cross section ilnes and are counted twice," REFERENCES, 1,'U.'S. Nstlonal Health Survey.:~rlgln and Pro- ~ram of the U,-S, Nationa| Health Survey, Health Ststistles.- Series A*l,-Public Health Service Pub- llcation No.S84-Alo:Public Heal.th Service,:WashlnE- ton, D,C,, May 2. Hansen, M.H.; Hurwitz, W,N,; and Madow, Sample Survey Methods and Theory.Vo1,-I.-~ohn Wiley ~ Sons, Inc.~ New York, N,'Y,, 3, Cochran. W.G,~Sampllng Techniques. ~ohn Wiley • Sons, Inc,, New York, 4, Demlng. W,E,: Some Theory of Sampling, ~ohn Wiley & Sons~ In~0, New York, N.'Y.~ |950,~p, Tl06653854

REPORTS FROM THE U. S. NATIONAL HEALTH SURVEY Series A I. origin and Program of the g. S. NBtional Health Survey. Public Health Service PubIicatlon No. 584-AI. Price 2S cents. 2. The Statistical Design of the Health Household-lnterview Sur- vey. Public Health Service Publication No. 58~-A2. Series B I. Preliminary Report on Volume of Physician Visits, United States, July-September 1957. Public Health Service Publication No. 58~-B1. Price 25 cents. 2o Preliminary Report on Volume of Dental Care, United States, July-September 1957. Public Health Service Publ ~cation No, 58~-B2. Price 25 cents. 3. Preliminary Report on Number of Persons Injured, United States, July-December %95?, Public Health Service Publication No, 58~-B3, Price 30 cents. Preliminary Report on Disability, United states, July-September %967. Public Healt~ Service Publicatlon No. 58~-B~, Price 30 cents. The Library of Concjress Catalog Card The st:~tistieal design (ff the health }musehohl-iutervie~ survey, hy staff of the V. S. Nali(mnl lh,~dih Survey and the Bureau of the 0ensus. Washington, I'. S. DePt. of Health, l,Muca/i~)n, ,rid Well,r[', l'ubli~" Ileal/h Service, Division of Public Health Methods, tll I'. ilhl~., l~l~lp. 26 ~.m. ¢ltr ]Tealth stulisti(~, s,,r. A-2) ~l'. $4 Plfl)li~. Ih, alth Si,rvieo. Publication Ira. 7~4-A2. l{ibli(~grIIphy : p. 411. 1. Ilvulth surve3-s. RAIl.B15474 no.:t 614.U97:~ 58-t;1-Mt TI06653855

---