The data presented in this Chapter were taken from the Health Promotion and Dis- ease Prevention Supplement to the 1985 NHIS and the Cancer Control Supplement to the 1987 NHIS. National Health and Nutrition Examination Survey and Hispanic Health and Nutri- tion Examination Survey Since 1960, the National Center for Health Statistics has conducted periodic health surveys that have included physical examinations and laboratory tests. Initially called the National Health Examination Survey (NHES), the name of this survey was changed to the National Health and Nutrition Examination Survey (NHANES) in 1970 when a nutrition component was added. The NHES was conducted in 1960, 1963, and 1966, and the NHANES in 1971, 1976, and 1988. Although the NHANES as a population survey included all of the Nation's major subpopulations including Hispanics, the sample sizes were insufficient to produce reli- able estimates of health status, particularly if the three major Hispanic subgroups- Mexican-Americans, Cuban-Americans, and Puerto Ricans-were considered separately. Therefore, the Hispanic Health and Nutrition Examination Survey (HHANES) was developed by the National Center for Health Statistics. The HHANES was designed to provide sufficient samples of each Hispanic subgroup. The survey not only produces reliable estimates of health status for each subgroup but also permits cross-cultural comparisons within the broader Hispanic cultural context. The HHANES was a probability-based survey of three distinct subgroups of a major U.S. minority group rather than of a national sample. The sampling methodology used complex, multistaged, stratified, clustered samples of the defined population. When weighted, the sample data represent the targeted population. For HHANES, the tar- geted population consisted of three groups of civilian, noninstitutionalized persons, aged 6 months to 74 years from three areas of the country that had a sufficient number or proportion of Hispanics to render it economically feasible to screen households and to operate an examination center: (1) Mexican-Americans residing in selected areas of Texas, California, Colorado, New Mexico, and Arizona; (2) Cuban-Americans resid- ing in Dade County, Florida; and (3) Puerto Ricans residing in the New York City area. Data were collected from 1982 through 1984 via in-person household interviews and via examination at a local examination center. Information was collected regarding a number of health issues, including the use of tobacco. NIDA High School Seniors Surveys on Drug Use Each year since 1975, the Monitoring the Future project has conducted surveys of representative national samples of high school seniors in the United States (Johnston, O'Malley, Bachman 1987). Monitoring the Future is conducted by the University of Michigan Institute for Social Research and receives its core funding from the Nation- al Institute on Drug Abuse. Each year, a multistage sampling procedure is used to identify approximately 135 public and private schools (the number of private schools has varied from 14 to 22) that 251 represent an accurate cross-section of high school seniors throughout the coterminous United States. The first stage involves the use of 74 primary sampling units developed by the University of Michigan Survey Research Center for use in its nationwide inter- view surveys. The second sampling stage involves choice of a single high school from most geographic areas (more than one is chosen in major metropolitan areas). The prob- ability of selection of any school is proportional to the size of the senior class. When a sampled school is unwilling to participate, a replacement school is selected from the same geographic area. Response rate of schools has been from 66 to 80 percent throughout the survey period. Up to 400 seniors are surveyed from each school. In schools with more than 400 seniors, a random sampling system convenient for the school (provided it results in an unbiased sample) is used to choose the 400 students to be interviewed. Most schools use the classroom as the basis for this selection. The total number of students inter- viewed each year has been between 15,700 and 19,000. The student response rate has varied from 77 percent to 84 percent throughout the survey period. The questionnaire administration in each school is carried out by local Survey Re- search Center representatives and their assistants following standardized procedures detailed in a project manual. Questionnaires are generally delivered in classrooms during normal class periods, although in some instances larger groups are used. Be- cause of the range of topics, five different questionnaire forms are used in the survey. These are distributed to participants in an ordered sequence to produce identical sub- samples. All five forms contain core data on demographics and some drug use (about one-third of the form); all other questions are asked of subsamples of the total respon- dents. Basic questions on cigarette usage have been included in the core for all years. Followup surveys by mail are conducted annually using representative subsamples from each of the previously participating classes, that is, the classes of 1976 through 1987. Thus, long-term panel data are collected on individuals, and analyses aimed at separating secular, age, and cohort effects are possible. (See O'Malley, Bachman, Johnston 1988.) NIDA National Household Surveys on Drug Abuse NIDA conducted household surveys on drug use in 1979,1982, and 1985. Data were obtained from a stratified random sample of 8,OCIO U.S. households; approximately 2,000 in-person interviews were conducted with respondents in the 12- to 17-year-old age group. Questions included whether any cigarettes were smoked within 30 days as well as within the previous year. Roper Survey, 1978 This survey was conducted for the Tobacco Institute via face-to-face interviewing with 2,5 11 subjects. Other methodological details are unavailable. 252 Roper Survey, 1980 The 1980 Roper Survey used face-to-face interviews to test a nationally repre- sentative sample of 2,000 adults for knowledge about the health hazards of smoking. The study was commissioned by the FTC and was conducted in November 1980. The total sample was split into two halves, and one set of questions was varied between the two. Thus, the sample size for several of the questions on the health effects of smok- ing was approximately half the total sample size. US DHEW Teenage Smoking Surveys In 1968, 1970, 1972, 1974, and 1979, random samples of teenagers aged 12 to 18 years were surveyed by telephone in December-January (US DHEW 1972, 1976b, 1979b). The first stage of the 3-stage sampling plan involved grouping and selecting telephone exchanges and was designed to eliminate geographic bias. Within the selected exchanges, equal numbers of random-digit-dialed telephone numbers were generated and contacted. Household enumeration was undertaken with an adult respon- dent and if more than one person aged between 12 and 18 years lived in the house, ran- dom selection was used to choose the study participant. In 1968, the sample size was 4,931, 89 percent of whom were interviewed by telephone. The other 11 percent lived in nontelephone households and were interviewed in their homes. As exclusion of the nontelephone households did not substantially af- fect prevalence estimates, later surveys did not include household interviewing of non- telephone households. The sample size in 1970 was 2,640; in 1972. it was 2,790; in 1974, it was 2,553; and in 1979, it was 2,639. In 1979, a followup survey was also un- dertaken of 1,194 (46.8 percent) of the 1974 respondents. Approximately 12,ooO households were contacted in 1979, from which 2,639 people aged 12 to 18 years were interviewed. In no survey was there any attempt to validate the smoking status indi- cated. 253 References ALLPORT, G.W. Attitudes. In: Murchison, C. (ed.) A Handbook ofSocial Psychology. War- tester, Massachusetts: Clark University Press, 1935. AMERICAN ASSOCIATION FOR RESPIRATORY CARE. Majority of airline passengers want smoking ban; National health group calls for action. Press release. Dallas: American Association for Respiratory Care, September 1, 1987. AMERICAN CANCER SOCIETY. Release of the Lieberman Survey. Press release. New York: American Cancer Society. November 10, 1986. AMERICAN CANCER SOCIETY. Cancer Facts and Figures--1988. New York: American Cancer Society, 1988a. AMERICAN CANCER SOCIETY. The 1987 Survey of Public Awareness and Use of Cancer Tests. Sponsored by the American Cancer Society, conducted by the Gallup Organization, analyzed by Lieberman Research, 1988b. AMERICAN LUNG ASSOCIATION. Americans want smoke-free air at work. Press release. New York: American Lung Association, December 5, 1985. AMERICAN LUNG ASSOCIATION. Smokers know risks but don't say "no" to cigarettes. Press release. New York: American Lung Association, December 4, 1987. AMERICAN MEDICAL ASSOCIATION. Americans favor tobacco advertising ban, says AMA survey. Press release. Chicago: American Medical Association, October 29, 1986. BACHMAN, J.G., JOHNSTON, L.D., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1976. AM Arbor, Michigan: University of Michigan, 1980a. BACHMAN, J.G., JOHNSTON, L.D., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1978. AM Arbor, Michigan: University of Michigan, 1980b. BACHMAN, J.G., JOHNSTON, L.D., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1980. Ann Arbor, Michigan: University of Michigan, 1981. BACHMAN, J.G., JOHNSTON, L.D., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1982. Ann Arbor, Michigan: University of Michigan, 1984. BACHMAN, J.G., JOHNSTON, L.D., O'MALLEY. P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1984. Ann Arbor, Michigan: University of Michigan, 1985. BACHMAN, J.G.. JOHNSTON, L.D., O'MALLEY, P.M. Monitoring fhe Fufure. Question- naire Responses From rhe Nation's High School Seniors, 1986. Ann Arbor, Michigan: University of Michigan, 1987. BECKER, D.M., LEVINE, D.M. Risk perception, knowledge, and lifestyles in siblings of people with premature coronary disease. American Journal of Preventive Medicine 3( 1):45- 50, 1987. BEN-SIRA, 2. Health promoting function of mass media and reference groups: Motivating or reinforcing of behavior change. Social Science and Medicine 16(7):825-834, 1982. CANADIAN GALLUP POLL, October 1986. CENTERS FOR DISEASE CONTROL. Survey to assess the prevalence, attitudes, knowledge, and beliefs about smoking behavior among the adult population. Survey questionnaires. Prepared by Westat for the Office on Smoking and Health, Rockville, Maryland, August 1986. CENTERS FOR DISEASE CONTROL. Passive smoking: Beliefs, attitudes, and exposures- United States, 1986. Morbidity and Mortality Weekly Report 37( 15):239-241,April22,1988. CHILTON RESEARCH SERVICES. A Survey of Adolescent and Adult Attitudes, Values, Be- havior, Intentions and Knowledge Related to Cigarette Smoking. Final Report. Submitted to the Federal Trade Commission, June 25, 1980. 254 DOLECEK, T.A., SCHOENBERGER, J.A., OMAN, J.K., KREMER, B.K., SUNSERI, A.J., ALBERTI, J.M. Cardiovascular risk factor knowledge and belief in prevention among adults in Chicago. American Journal of Preventive Medicine 2(5):262-267, 1986. FEDERAL TRADE COMMISSION. Federal Trade Commission StaffReport on the Cigarette Smoking Advertising Investigation. Federal Trade Commission, May 1981. FISHBEIN, M. Consumer Beliefs and Behavior With Respect to Cigarette Smoking: A Criti- cal Analysis of the Public Literature. A Report Prepared for the Staff of the Federal Trade Commission. Federal Trade Commission, May 1977. FOLSOM, A.R., SPRAFKA, M., LUEPKER, R.V., JACOBS, D.R. JR. Beliefs among black and white adults about causes and prevention of cardiovascular disease: The Minnesota Heart Study. American Journal of Preventive Medicine 4(3): 121-127. 1988. FOX, S.H., BROWN, C., KOGNTZ, A.M., KESSEL, S.S. Perceptions of risks of smoking and heavy drinking during pregnancy: 1985 NHIS findings. Public Health Reports 102(1):73- 79, January-February 1987. GALLUP. Smoking in America. Public Attitudes and Behavior. Gallup Report No. 155. Princeton, New Jersey: Gallup Organization, June 1978. GALLUP. Gallup Smoking Audit. The Gallup Report. Report No. 190. Princeton, New Jer- sey: Gallup Organization, July 1981. GALLUP. Survey of attitudes toward smoking. Conducted for the American Lung Associa- tion. Princeton, New Jersey: Gallup Organization, July 1985. GALLUP. Majority Backs Ban on Smoking in Public Places. Gallup Report No. 258. Prince- ton, New Jersey: Gallup Organization, March 1987a. GALLUP. Majority backs ban on smoking in public places. Press release. Princeton, New Jer- sey: Gallup Organization, April 3, 1987b. GALLUP. Survey of attitudes toward smoking. Conducted for the American Lung Associa- tion. Princeton, New Jersey: Gallup Organization, 1987~. GALLUP. Attitudes toward smoking in restaurants and fast food establishments. Conducted for the National Restaurant Association. Princeton, New Jersey: Gallup Organization, February 1987d. GALLUP. On-the-go Americans prefer smoke-free air. American Journal of Public Health 78(5):563, May 1988a. GALLUP. Majorities now support ban on public smoking, cigarette ads. Press release. Prince- ton, New Jersey: Gallup Organization, September 4, 1988b. HARRIS, L. Prevention in America: Steps People Take-Or Fail to Take-for Better Health. Louis Harris and Associates, 1983. HARRIS, L. The Prevention Index. 1988. Louis Harris and Associates, May 13. 1988. HARVEY, L., SHUBAT, S. Public opinion on health care issues, 1986. In: AMA Surveys of Physician andPublic Opinion, 1986. Chicago: American Medical Association, October 1986. HARVEY, L., SHUBAT, S. Public opinion on healthcare issues. In: AMASurveysofPhysiciun and Public Opinion on Health Care Issues, 1987. Chicago: American Medical Association, April 1987. JOHNSTON, L.D., BACHMAN, J.G. Monitoring the Future. Questionnaire Responses From the Nation's High School Seniors, 1975. Ann Arbor, Michigan: University of Michigan, 1980. JOHNSTON, L.D., BACHMAN, J.G., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1977. AM Arbor, Michigan: University of Michigan, 1980a. JOHNSTON, L.D., BACHMAN, J.G., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1979. Ann Arbor, Michigan: University of Michigan, 198Ob. 255 JOHNSTON, L.D., BACHMAN, J.G., O'MALLEY. P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1981. Ann Arbor, Michigan: University of Michigan, 1982. JOHNSTON, L.D., BACHMAN, J.G., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1983. Ann Arbor, Michigan: University of Michigan, 1984. JOHNSTON, L.D., BACHMAN, J.G., O'MALLEY, P.M. Monitoring the Future. Question- naire Responses From the Nation's High School Seniors, 1985. Ann Arbor, Michigan: University of Michigan, 1986. JOHNSTON, L.D., O'MALLEY, P.M., BACHMAN, J.G. National Trends in Drug Use and Related Factors Among High School Students and Young Adults. 1975-1986. U.S. Depart- ment of Health and Human Services, Public Health Service, Alcohol, Drug Abuse, and Men- tal Health Administration, National Institute on Drug Abuse. DHHS Publication No. (ADM) 87-1535, 1987. LEVENTHAL, H., GLYNN, K., FLEMING, R. Is the smoking decision an `informed choice'? Effect of smoking risk factors on smoking beliefs. Journal of the American MedicalAssocia- tion 257(24):3373-3376, June 26, 1987. LIEBERMAN RESEARCH. A Study of Public Attitudes Toward Cigarette Advertising and Promotion Programs--Conducted for American Cancer Society, American Heart Associa- tion, American Lung Association. Lieberman Research, October 1986. MAHONEY, M.J. Cognition and BehaGor Modification. Cambridge, Massachusetts: Bal- linger Publishing, 1974. MASSEY, M.M., BOYD, E., MATTSON, M., FEINLEIB, M.R. Inventory of surveys on smok- ing. Public Health Reports 102(4):43@438, July-August 1987. MATTSON, M.E., POLLACK, ES., CULLEN, J.W. What are the odds that smoking will kill you? American Journal of Public Health 77(4):42543 I, April 1987. NATIONAL CENTER FOR HEALTH STATISTICS, THORNBERRY, O.T., WILSON, R.W., GOLDEN, P.M. Health promotion data for the 1990 objectives, estimates from the National Health Interview Survey of Health Promotion and Disease Prevention, United States, 1985. Advance Data From Vital and Health Statistics, No. 126. U.S. Department of Health and Human Services, Public Health Service. DHHS Publication No. (PHS) 86-1250, September 19, 1986. NATIONAL CENTER FOR HEALTH STATISTICS. THORNBERRY, 0. An experimental comparison of telephone and personal health interview surveys. Vita/ and Health Statistics, Series 2, No. 106. U.S. Department of Health and Human Services, Public Health Service. DHHS Publication No. (PHS) 87-1380, August 1987. NATIONAL CENTER FOR HEALTH STATISTICS, SCHOENBORN, C.A. Health promo- tion and disease prevention: United States, 1985. Vita/and Health Statistics, Series IO, NO. 163. Public Health Service. DHHS Publication No. (PHS) 88-1591, February 1988. O'MALLEY, P.M., BACHMAN, J.G., JOHNSTON, L.D. Period, age, and cohort effects on substance use among young Americans: A decade of change, 1976-86. American Journal of Public Health 78( LO): 13 15-l 32 I, October 1988. OSKAMP, S. Attitudes and Opinions. Englewood Cliffs. New Jersey: Prentice-Hall. 1977. PERLSTADT, H., HOLMES, R.E. The role of public opinion polling in health legislation. American Journal of Public Health 77(5):612-614, May 1987. REMINGTON, P.L., FORMAN, M.R., GENTRY, E.M., MARKS, J.S., HOGELIN, G.C., TROWBRIDGE, F.L. Current smoking trends in the United States. The 1981-1983 Be- havioral Risk Factor Surveys. Journal of the American Medical Association 253(20):2975- 2978, May 24-31, 1985. ROPER, A Study of Public Attitudes Toward Cigarette Smoking and the Tobacco Industy in 1978, Volume I. Roper Organization, May 1978. 256 SCHINKE, S.P., GILCHRIST, L.D. Preventing cigarette smoking with youth. Journal of Primaty Prevention 5( 1):48-56, Fall 1984. SHOPLAND, D.R.. BROWN, C. Toward the 1990 objectives for smoking: Measuring the progress with 1985 NHIS data. PuhlicHealthReports 102( 1):68-73, January-February 1987. SLOVIC, P. Informing and educating the public about risk. Risk Analysis 6(4), 1986. STEWART, P.J., DUNKLEY, G.C. Smoking and health care patterns among pregnant women. Canadian Medical Association Journal 133( 10):989-994, November 15, 1985. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Smoking for Women. A Report of the Surgeon General. U.S. Department of Health and Human Services, Public Health Service, Office of the Assistant Secretary for Health, Office on Smoking and Health, 1980. US. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Smoking. The Changing Cigarette: A Report of the Surgeon General. U.S. Department of Health and Human Services, Public Health Service, Office of the Assistant Secretary for Health, Office on Smoking and Health. DHHS Publication No. (PHS) 8 I -50156, 198 I. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Smoking: Cancer. A Report of the Surgeon Genera/. U.S. Department of Health and Human Services. Public Health Service, Office on Smoking and Health. DHHS Publication No. (PHS) 82-50179, 1982. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Smoking: Cardiovascular Disease. A Report of the Surgeon General. U.S. Department of Health and Human Services, Public Health Service, Office on Smoking and Health. DHHS Publication No. (PHS) 84-50204, 1983. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Smoking: Chronic Obstructive Lung Disease. A Report of the Surgeon General. U.S. Depart- ment of Health and Human Services, Public Health Service, Office on Smoking and Health. DHHS Publication No. (PHS) 84-50205, 1984. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Smoking: Cancer and Chronic Lung Disease in the Workplace. A Report of the Surgeon Genera/. US. Department of Health and Human Services, Public Health Service, Office on Smoking and Health. DHHS Publication No. (PHS) 85-50207, 1985. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Involuntary Smoking. A Report of the Surgeon General. U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control. DHHS Publication No. (CDC) 87-8398, 1986a. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The 1990 Health Objectives for the Nation: A MidcourseReview. U.S. Department of Health and Human Services, Public Health Service, Office of Disease Prevention and Health Promotion, November 1986b. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Using Smokeless Tobacco. A Report of the Advisory Committee to the Surgeon General, 1986. U.S. Department of Health and Human Services, Public Health Service. NIH Publication No. 86-2874, April 1986~. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. Youth Use of Smokeless Tobacco: More Than a Pinch of Trouble. U.S. Department of Health and Human Services, Office of Inspector General, Office of Analysis and Inspections, January 1986d. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Consequences of Smoking: Nicotine Addiction. A Report of the Surgeon General, 1988. U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Center for Health Promotion and Education, Office on Smoking and Health. DHHS Publication No. (CDC) 88-8406,1988. 257 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. Adult Use of Tobacco, 1986. U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Office on Smoking and Health, in press. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. Use of Tobacco. prac. rices, Attitudes, Knowledge, and Beliefs. United States-Fall 1964 and Spring 1966. U.S. Department of Health, Education, and Welfare, Public Health Service, Health Services and Mental Health Administration, National Clearinghouse for Smoking and Health, July 1969. U.S. DEPARTMENTOFHEALTH, EDUCATION, AND WELFARE. Teenage Smoking. Nu- tional Patterns of Cigarette Smoking, Ages 12 Through 18, in 1968 and 1970. U.S. Depart- ment of Health, Education, and Welfare, Health Services and Mental Health Administration, National Clearinghouse for Smoking and Health. DHEW Publication No. (HSM) 72-7508, 1972. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. Adult Use of Tobacco, 1970. U.S. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control, National Clearinghouse for Smoking and Health, June 1973. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. Cigarette Smoking Among Teen-agers and Young Women. U.S. Department of Health, Education, and Welfare, Public Health Service, National Institutes of Health. DHEW Publication No. (NIH) 77-1203, 1975a. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. The Health Conse- quences of Smoking. 1975. U.S. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control. HEW Publication No. (CDC) 778704, 1975b. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. Adult Use of Tobac- co-1975. U.S. Department of Health, Education, and Welfare, Public Health Service, Cen- ter for Disease Control, 1976a. U.S. DEPARTMENTOF HEALTH, EDUCATION, AND WELFARE. TeenageSmoking: Na- tional Patterns of Cigarette Smoking, Ages 12 Through 18, in 1972 and 1974. U.S. Depart- ment of Health, Education, and Welfare, Public Health Service, National Institutes of Health. DHEW Publication No. (NIH) 76-93 1, 1976b. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. The Health Conse- quences of Smoking, 1977-1978. U.S. Department of Health, Education, and Welfare, Public Health Service, Office of the Assistant Secretary for Health, Office on Smoking and Health. DHEW Publication No. (PHS) 79-50065, 1978. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. Smoking and Health. A Report of the Surgeon General. U.S. Department of Health, Education, and Welfare, Public Health Service, Office of the Assistant Secretary for Health, Office on Smoking and Health. DHEW Publication No. (PHS) 79-50066, 1979a. U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE. Teenage Smoking. Im- mediate and Long-Term Patterns. U.S. Department of Health, Education, and Welfare, Na- tional Institute of Education, November 1979b. U.S. PUBLIC HEALTH SERVICE. Smoking and Health. Report of the Advisory Committee to the Surgeon General of the Public Health Service. U.S. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control. PHS Publication No. 1103, 1964. U.S. PUBLIC HEALTH SERVICE. The Health Consequences of Smoking. 1969 Supplement to the 1967 Public Health Service Review. U.S. Department of Health, Education, and Wel- fare, Public Health Service. PHS Publication No. 1696-2 (Supplement), 1969. WEEKLY READER. Weekly Reader National Survey on Safety and Health. Middletown, Con- necticut: Field Publications, 1986. 258 CHAPTER 5 CHANGES IN SMOKING BEHAVIOR AND KNOWLEDGE ABOUT DETERMINANTS 259 CONTENTS Introduction ....................................................... 263 Part I. Changes in Smoking Behavior .................................. 263 Trends in Cigarette Smoking ....................................... 263 Introduction ............................................... 263 Nature and Quality of Data ................................... 264 Excise Tax and Sales Data ................................ 264 Self-Reported Survey Data ................................ 264 Validity of Self-Reported Survey Data ....................... 265 Correlation Between Self-Reported Survey Data and Sales Data ........................................... 265 Trends in Cigarette Sales ..................................... 267 Trends in Smoking Prevalence Among Adults .................... 267 Cigarette Smoking by Sex, Race (Whites and Blacks), and Educational Attainment (National Health Interview Surveys: 1965-87) .... 267 Adult Use of Tobacco Surveys: 1964-86 .................... 27 1 Cigarette Smoking Among Different Occupational Groups ...... 272 Special Populations: Hispanics ............................ 274 Special Populations: American Indians and Alaskan Natives ..... 274 Special Populations: Asian Americans ...................... 275 Special Populations: Pregnant Women ...................... 275 Special Populations: Military Personnel ..................... 276 State-Specific Smoking Prevalence ......................... 278 Behavioral Risk Factor Surveillance System: 1982-87 ...... 278 Current Population Survey: 1985 ....................... 279 BRFS and CPS Comparison ........................... 28 1 Summary .............................................. 284 Trends in Quitting ........................................... 285 Introduction .......................................... ..28 5 Trends in the Proportion of Smokers Quitting (Quit Ratio) (NHIS) 286 Smoking Continuum (AUTS) .............................. 286 Summary ............................................ ..29 2 Trends in the Proportion of Smokers Who Are Heavy Smokers ....... 292 Trends in the Initiation of Smoking ............................. 296 Age of Initiation ........................................ 296 Prevalence in 20- to 24-Year Age Group ..................... 298 Trends in Adolescent Smoking ................................ 302 NIDA High School Seniors Surveys on Drug Use, 1976-87 ...... 302 1987 National Adolescent Student Health Survey .............. 306 US DHEW Teenage Smoking Surveys, 1968-79 ............... 307 NIDA National Household Surveys on Drug Abuse, 1979-85 .... 312 Summary ............................................ ..313 Changes in the Types of Cigarettes Smoked ...................... 3 13 Filtered Cigarettes ....................................... 3 13 261 Low-Tar, Low-Nicotine Cigarettes .......................... 313 Menthol Cigarettes ...................................... 3 17 CigaretteLength ...................................... ..3i 7 Summary and Comment .................................. 318 Other Types of Tobacco Use ....................................... 3 19 Smokeless Tobacco Use ..................................... 3 19 Cigar and Pipe Smoking ..................................... 322 Part II. Changes in Knowledge About the Determinants of Smoking Behavior , 329 Introduction: Historical and Conceptual Overview ...................... 329 1964 Surgeon General's Report ................................ 329 1979 Surgeon General's Report ................................ 330 CurrentViews ............................................ ..33 1 Development of Smoking .......................................... 333 Pharmacologic Processes and Conditioning ...................... 333 Cognition and Decisionmaking ................................ 335 Personal Characteristics and Social Context ...................... 336 Personal Characteristics .................................. 336 Peer and Family Influences ................................ 336 Cigarette Marketing ..................................... 338 Summary ................................................ ..33 9 RegularSmoking ............................................... ..34 0 Pharmacologic Processes and Conditioning ...................... 340 Nicotine Addiction ...................................... 340 Mechanisms of Nicotine Action ............................ 342 Conditioning and Smoking ................................ 343 Cognition and Decisionmaking ................................ 344 Personal Characteristics and Social Context ...................... 345 Personal Characteristics .................................. 345 Social Context Influences ................................. 347 Summary ................................................ ..34 8 Cessation and Relapse ............................................. 348 Pharmacologic Processes and Conditioning ...................... 349 Cognition and Decisionmaking ................................ 350 Expectancy-Value Models ................................ 350 Self-EfficacyandSmoking .............................. ..35 1 Outcome Expectations ................................... 352 Personal Characteristics and Social Context ...................... 353 Personal Characteristics .................................. 353 Social Context Influences ................................. 354 Summary .................................................. 355 Summary of Changes in Knowledge About Determinants of Smoking Behavior356 Conclusions ....................................................... 357 Appendix ....................................................... ..35 9 References ...................................................... ..36 1 262 INTRODUCTION This Chapter reviews two major aspects of smoking behavior since release of the first Surgeon General's Report on smoking and health in 1964: (1) changes in smoking be- havior in the United States (Part I) and (2) changes in our knowledge about the deter- minants of smoking during this period (Part II). During the past 25 years, the prevalence of cigarette smoking has declined in virtual- ly every major so&demographic group, including men and women, adults and adoles- cents, blacks and whites, and persons with and without college education. This decline has been particularly evident among men, in whom the prevalence of smoking declined from 50 percent in 1965 to 32 percent in 1987. The first part of this Chapter analyzes trends in smoking prevalence, cessation, and initiation, and examines smoking patterns among different so&demographic groups and other special populations. These analyses are based, for the most part, on cross-sectional population-based data collected periodically since 1964. At the same time, our knowledge about determinants of smoking has increased sub- stantially. Physiological, behavioral, and social factors that may influence the initia- tion and maintenance of smoking have been extensively researched. Many important predictors of initiation, quitting, and relapse have been identified. The development of this body of knowledge is reviewed in the second part of this Chapter. Information reviewed in that part of the Chapter is primarily derived from research studies and in- tervention trials that employ smaller sample sizes than the population-based surveys used in Part I. These studies, however, usually collect more detailed information and often obtain longitudinal followup data. PART I. CHANGES IN SMOKING BEHAVIOR Trends in Cigarette Smoking Introduction Accurate information on trends in smoking prevalence in the major so&demographic groups in the United States is of interest to public health officials, policymakers, researchers, clinicians, and news media. These data are important for estimating the magnitude of the problem of smoking and for targeting public health in- terventions to those at highest risk of smoking. Accurate data on trends in smoking (including initiation and quitting) are necessary to be able to project future smoking patterns. Accurate projections must be available, in turn, to set appropriate but realistic goals for key future years (e.g., 1990,200O). This Section analyzes trends in smoking prevalence, quitting, and initiation during the past quarter century. Data on smoking prevalence in the 1940s and 1950s from Gallup sur- veys and the Current Population Survey have been cited elsewhere (CDC 1987a; US DHHS 1988, Appendix A). 263 Changes in measures of smoking behavior (e.g., prevalence, quitting, initiation), like any quantitative variables, can be calculated as absolute or relative changes. For than- ges in percentages, the absolute change would be in percentage points; the relative (per- cent) change would be calculated by subtracting the "new" percentage from the base percentage, dividing the difference by the base percentage, and multiplying the quotient by 100. Each measure of change has advantages and disadvantages. Throughout Part 1 of this Chapter, changes in smoking prevalence, quitting, and initiation are described primarily in terms of absolute changes. Nature and Quality of Data A number of sources of information provide insight into smoking behavior in the United States. These sources fall into two main categories: those based on excise taxa- tion of cigarettes and those based on population surveys of self-reported smoking. Excise Tax and Sales Data The Economic Research Service of the U.S. Department of Agriculture (USDA) has estimated total and adult per capita consumption of cigarettes for a number of years. These estimates are based on data from the Bureau of Alcohol, Tobacco and Firearms (Department of Treasury), the Bureau of Commerce (Department of Commerce), the Tobacco Institute, and other private and industry sources. The Tobacco Institute reports the number of packs of cigarettes on which State taxes are paid; the Bureau of Alcohol, Tobacco and Firearms reports the number of cigarettes on which Federal taxes are paid; and the Bureau of Commerce reports the number of cigarettes imported into the United States. Both Federal and State taxes are excise taxes collected at the wholesale level (on removals) and are not standard sales taxes. The estimated level of consumption is based on both Federal and State taxes on removals, as well as on imports, and is adjusted for estimated inventory changes. Adult per capita consumption is customarily calculated in the United States by dividing total consumption by the total estimated population 18 years of age and older. (The World Health Organization (1988) has published per capita cigarette consumption figures for countries throughout the world based on the population 15 years of age and older.) Self-Reported Survey Data A number of different data sources are available to assess national trends in smoking during the past 25 years. These surveys differ on the basis of sample size, method of data collection (telephone interview versus face-to-face household interview versus questionnaire administered in school), population (adults versus adolescents), sampling frame (national versus State based), and the extent of information collected on tobac- co use. Details of the methodology for the various surveys are provided in the Appen- dix to Chapter 4 and in Table 1 of that chapter. The amount of information provided varies from survey to survey depending on the availability of information. 264 Validity of Self-Reported Survey Data The validity of self-reports of smoking status from surveys may affect the usefulness Jf these data in reporting historical trends. Respondents' sensitivity to the social stig- ma associated with smoking has been cited as a reason persons might underreport their smoking status (Warner 1978; Kozlowski 1986). Whereas biochemical assessment is generally more reliable than self-report in assessing level of nicotine intake (US DHHS 1988), self-reported data appear valid for estimating prevalence of smoking in the population. Forexample, studies of patients in several settings (Petitti, Friedman, Kahn 1981; Pojer et al. 1984), as well as two large community studies (Fortmann et al. 1984; Pierce, Dwyer et al.1 987b),have shown that measurement of smoking by self-report or by biochemical markers gives approximately the same estimates of prevalence. A more recent study of 1,3 17 Hispanics, however, showed that self-reported cigarette use un- derestimated biochemically validated use (Coultas et al. 1988). It is possible that the accuracy of self-reported data will vary depending on whether the data collection method is face to face or by telephone interview. Although biochemical-validation data do not exist to allow the quantification of such a difference, comparisons of smoking prevalence estimates derived from surveys using telephone versus in-person interviews have shown that the former are generally 1 to 3 percentage points below the latter (CDC 1987a; see below and NCHS 1987). In addition, concerns have been expressed about the validity of data reported by one person on behalf of another ("proxy response") (NCHS 1985. p. 54). For adults, these concerns relate more to measures of the number of cigarettes smoked per day than to the classification of whether a person is a current smoker (US DHEW 1969, p. 794; Rogot and Reid 1975; National Research Council 1986, pp. 110-l 12). For adolescents, proxy reporting may also affect prevalence estimates (Millar 1985). Correlation Between Self-Reported Survey Data and Sales Data Warner (1978) compared self-reported data on cigarette consumption with USDA consumption data for the years 1964-75. He found that self-reported cigarette con- sumption increasingly underestimated the USDA estimates, possibly because of the in- creasing social stigma associated with smoking. Changing social acceptability of smoking would not be expected to affect the USDA estimates. To the extent that a "so- cial acceptability" bias in self-reported data may have increased in recent years, the dramatic decrease in smoking prevalence observed during the past 25 years could be in part artifactual. Hatziandreu et al. (in press) analyzed more recent data to determine whether the trend reported by Warner (1978) has continued. Self-reported consumption data for adults and teenagers were obtained from the National Health Interview Survey (NHIS) (Na- tional Center for Health Statistics (NCHS)) and the National Household Survey on Drug Abuse (National Institute on Drug Abuse (NIDA)). Self-reported cigarette consump- tion was estimated based on the smoking prevalence, the average self-reported number of cigarettes smoked per day, and the U.S. population size each year. A "consumption ratio" was calculated by dividing self-reported consumption by USDA estimates ob- 265 tained from cigarette tax data. This ratio has been relatively stable recently, varying from 0.73 in 1974 to 0.69 in 1976 with a mean of 0.72 (Table 1). A least-squares regres- sion analysis was used to identify any trend. The slope of the regression line was not significantly different from zero (p=O.85), countering the hypothesis that self-reported data are increasingly underestimating actual cigarette consumption. These results sug- gest that national surveys provide a reliable estimate of U.S. smoking trends. The reasons for the consistent difference between cigarette consumption based on excise tax data versus self-reported data are unclear; one possible explanation would be a sys- tematic bias from "rounding down" of self-reported daily consumption to the nearest multiple of a half-pack (see Table 14 and related discussion and Kozlowski 1986). TABLE l.-Estimates of cigarette consumption in the United States, based on cigarette excise taxes and self-reports, M4-85 Year 1974 1976 1978 1979 1980 1983 1985 Excise taxes Self-reported (billions) (billions) 599.0 434.9 613.5 424.4 616.0 438.4 621.5 441.2 631.5 459.1 600.0 487.8 594.0 414.4 Fraction 0.73 0.69 0.71 0.71 0.73 0.78 0.70 NOTE: Estimated by the U.S. Department of Agriculture. Self-reported consumption includes estimatedconsumption for adults (NHIS. NCHS) and estimated consumption for adolescents (National Household Survey on Drug Abuse. NIDA). SOURCE: Hatziandreu et al.. in press. The difference in the findings reported by Hatziandreu et al. (in press) and Warner (1978) may relate to differences in methodology. For example, Warner used data from the 1964, 1966, 1970, and 1975 Adult Use of Tobacco Surveys (AUTSs). He found that the major decrease in the consumption ratio occurred between 1966 and 1970. This may have occurred because the 1964 and 1966 AUTSs were in-person surveys, whereas the 1970 and 1975 AUTSs were telephone surveys. As mentioned above, telephone surveys generally provide slightly lower estimates of smoking prevalence than in-per- son surveys. On the other hand, Hatziandreu et al. (in press) used only in-person inter- view data (NHIS) for adults and the NIDA Household Interview Survey on Drug Use for adolescents. The consumption ratios obtained by Warner for 1964 and 1966 (0.73 and 0.72, respectively) using in-person survey data were similar to the mean ratio (0.72) reported by Hatziandreu et al. for the period 1974-85. In addition, the 1974 in-person estimate was 0.73 (Hatziandreu et al., in press), whereas the 1975 telephone estimate was 0.64 (Warner 1978). This difference provides further evidence that the decrease in the consumption ratio reported by Warner was an artifact of the change in the AUTS methodology. 266 Trends in Cigarette Sales Total cigarette consumption in the United States (as estimated by sales data) increased steadily from 1900 until 198 1, when an estimated total of 640 billion cigarettes were smoked (Table 2). Since 198 1, there has been a steady decline in consumption despite increasing population size. The number of cigarettes smoked in 1987 is estimated at 574 billion. These figures refer to manufactured cigarettes and do not include roll-your-own cigarettes. Roll-your-own cigarettes have accounted for a declining proportion of total cigarettes consumed through the 20th century. By 1950, the estimated per capita con- sumption of roll-your-own cigarettes was 126, or 3.4 percent of total cigarettes con- sumed; in 1987, these figures were 23 and 0.7 percent, respectively (USDA, un- published data). Cigarette consumption data are divided by the population of adults 18 years of age and older to give an estimate of adult per capita consumption. This estimate represents the average number of cigarettes sold per adult in the population, not per smoker. It should be noted that trends in adult per capita consumption are somewhat biased be- cause there has been a trend over time for more people to start smoking regularly under age 18 (see section below on Trends in the Initiation of Smoking). Per capita consumption of manufactured cigarettes increased dramatically from its level of 54 cigarettes in 1900 to 4,171 cigarettes in 1960 (Table 2). From 1960-73, this figure remained relatively stable (compared with the previous rates of change) at about 4,000 cigarettes per year. Since 1973, there has been a yearly decline in per capita con- sumption. From 1973-87, this figure fell more than 23 percent to 3,196 cigarettes per year. Although there has been a decline in every one of these 15 years, the rate of decline has varied. From 1974-79, the magnitude of the yearly change increased rapid- ly until it reached a 2-percent decrease per year. In the 10 years since 1979, this decrease has fluctuated with a mean of 2.4 percent per year (standard deviation (S.D.) = 1.9). The large drop from 1982-83 (7.2 percent) was more than two standard deviations above the mean and is thought to be related, to a significant degree, to the March 1983 increase in the Federal cigarette excise tax from 8 cents per pack to 16 cents per pack (see Chapter 7). Trends in cigarette sales are also presented in Chapter 8 (Figure 3). Trends in Smoking Prevalence Among Adults Cigarette Smoking by Sex, Race (Whites and Blacks), and Educational Attainment (National Health Interview Surveys: 1965-87) Table 3 presents smoking prevalence from NHIS data for the years 1965,1966,1970, 1974, 197-O inclusive, 1983, and 1985, and preliminary data for 1987. These data are presented for the total adult population (aged 20 years and older) and by sex, race (whites and blacks), and educational attainment. They differ slightly from estimates published by NCHS (NCHS 1988c) because the data presented here are adjusted to the 267 TABLE 2.Total manufactured US. cigarette consumption and per capita consumption, adults aged 18 years and older, 1900-87 Year 1900 1910 I920 1930 1940 1950 Total consumption Per capita (billions) consumption 2.5 54 8.6 151 44.6 665 119.3 1.185 181.9 1.916 369.8 3,552 Percentage change in per capita consumption from previous year +IO.ga +16.@ +5.# +5.2a +6.0a I960 484.4 4,171 +1.f? 1961 502.5 4,266 +2.3 1962 508.4 4,266 0 I%3 523.9 4,345 +1.9 1964 511.3 4,194 -3.5 1965 528.8 4,258 +1.5 1966 541.3 4,287 +0.7 1967 549.3 4,280 -0.2 I%8 545.6 4.186 -2.2 1969 528.0 3,993 -4.6 1970 536.5 3,985 -0.2 1971 555.2 4,037 +I.3 1972 566.8 4.043 +o.l 1973 589.7 4.148 43.0 1974 599.0 4,141 -0.2 1975 607.2 4.123 -0.4 1976 613.5 4.092 -0.8 1977 617.0 4.05 1 -1.0 1978 616.0 3,967 -2.1 1979 621.5 3.861 -2.1 1980 631.5 3,844 xl.4 1981 640.0 3.836 -0.2 1982 634.0 3,739 -2.6 1983 600.0 3,488 -7.2 1984 600.4 3.446 -1.2 1985 594.0 3,370 -2.3 1986 583.8 3,274 -2.9 1987 (estimate) 574.0 3,196 -2.4 `Annualized rate of change during preceding decade. SOURCE: USDA (1987). 268 TABLE 3.-Trends in smoking prevalence (%), NHISs, United States, 196547, adults aged 20 years and older Sex Race Educational level Overall Less than High Some Y&W population Males Females Whites Blacks high school school graduate graduate college College graduate 1965' 40.4 1966 40.7 1970 37.0 1974 36.9 1976 36.1 1977 35.6 1978 34.0 1979 33.5 1980 33.3 1983 31.8 1985 30.4 1987b 29.1 Trend information (1965-85) Change'/year -0.50 Standard error 0.03 R= 0.97 50.2 31.9 40.0 43.0 50.8 32.0 40.4 42.9 36.5 41.1 42.5 33.7 44.3 30.8 36.5 41.4 34.8 38.3 36.1 28.1 43.4 31.4 36. I 44.0 36.5 31.6 36.9 28.3 42. I 31.3 35.6 41.2 35.8 37.8 36.4 27.4 40.9 31.4 34.9 41.8 35.8 38.4 35.2 25.6 39.0 29.6 33.6 38.2 35.3 36.5 32.1 23.8 38.4 29.2 33.2 36.8 34.9 35.4 33.3 23.4 38.5 29.0 32.9 31.2 35.5 35.7 31.2 24.6 35.5 28.7 31.4 36.6 34.7 35.6 30.0 19.9 33.2 28.0 29.9 36.0 35.7 34.2 28.1 18.4 31.7 26.8 28.8 34.0 35.7 33.1 26.1 16.3 -0.84 -0.21 -0.50 -0.39 -0.06 -0.32 -0.70 -0.76 0.04 0.03 0.03 0.08 0.03 0.05 0.07 0.08 0.98 0.81 0.97 0.74 NAd 0.87 0.94 0.93 `For 1%5. data stratified by education were not available. bPro"isicmal data only. %I percentage points. `?he slope of the regression line was not significantly different from zero. making the R* computation inappropriate SOURCE: NH& 1965-87: unpublished data, office on Smoking and Health. 1985 age distribution, whereas the previously published figures were adjusted to the 1970 age distribution. For each group, observed smoking prevalence for each survey year is reported. Ad- ditionally, to assess time trends from 1965-85, weighted least-squares regression analyses have been applied to these data. The 1987 data were not included in the regres- sion analyses because these data are preliminary estimates. These estimates can be used to provide a measure of predictive validity of the model; in general, the preliminary 1987 estimates are similar to projections from the model (Pierce, Fiore et al. 1989a). The R* statistic was used for each trend analysis and is a measure of how well the linear model fits the observed data values. R2 values may range from 0 (no linear trend) to 1.0 (a perfect fit between the observed values and a linear model). The data on overall smoking prevalence, as well as for each sex and racial group presented in Table 3, demonstrate linear trends with R2 values ranging from 0.74 to 0.98; thus, the models fit the data very well. Trends for three of the four educational categories are also fitted well by a linear model. For one category, less than high school graduation, no R* value is reported because the rate of change is very close to zero (making the R2 statistic inappropriate as an index of the amount of variation explained by the model). The change (in percentage points) per year is the slope of the line of best fit calculated by the model. The standard error of the slope allows confidence limits to be placed around the estimate of change per year. Ninety-five-percent con- fidence limits around the estimate of a slope are approximately equal to the slope plus or minus two times the standard error. Overall smoking prevalence declined from 40.4 percent in 1965 to 29.1 percent in 1987. The trend from 1965-85 is fitted almost exactly by a linear model (R2=0.97). Smoking prevalence in the United States adult population is decreasing at a rate of 0.50 percentage points per year with a standard error of 0.03. Thus, the 95-percent con- fidence interval for the change per year is 0.44 to 0.56. There is no evidence of any sudden deviations from the identified trend such as that seen in the per capita consump- tion data in 1983 (Table 2). The prevalence of smoking among men has decreased steadily from 50.2 percent in 1965 to 3 1.7 percent in 1987. The rate of decline between 1965 and 1985 was 0.84 per- centage points per year (95percent confidence limits, 0.76, 0.92). Female smoking prevalence remained stable at 31 to 32 percent from 1965-77. Subsequently, prevalence began to decline slowly and reached 26.8 percent in 1987. The overall rate of decline from 1965-85 was 0.21 percentage points per year (95percent confidence limits, 0.15,0.27). Fiore and coworkers (1989) have examined more recent trends in smoking by gender in greater detail. This analysis showed a rate of decline in prevalence among women of 0.33 percentage points per year between 1974 and 1985 (95~percent confidence limits, 0.21,0.45) (R*=O.SS). Although there has been a difference in smoking prevalence between blacks and whites, it may be explained by socioeconomic status (Novotny, Warner et al. 1988), and the rate of change in smoking prevalence in recent years has been similar between the races (Fiore et al. 1989). Smoking among whites decreased from 40.0 percent in 1965 to 28.8 percent in 1987. The rate of decline from 1965-85 was 0.50 percent- age points per year (95percent confidence limits, 0.44,0.56; R*=0.97). 270 For blacks the R* value for the simple linear model is 0.74, suggesting that the data should be reviewed more carefully. In 1965, 43.0 percent of blacks smoked. This number had changed little by 1977 when 41.8 percent smoked. From 1977-87, there was a considerable drop in smoking prevalence to 34.0 percent. Thus, the data suggest that there may be two trends among blacks. Fiore et al. (1989) fitted a linear model to the data for 1974-85 and reported a rate of change among blacks of -0.67 percent- age points per year with 95percent confidence limits of 0.37 and 0.97 (R*=O.SO). This rate of change was not significantly higher than that among whites for the same period (-0.57 percentage points per year). However, smoking prevalence among black men was decreasing at a faster rate than among white men ( 1.15 percentage points per year compared with 0.87, p=O.O3). There were no significant differences noted in the rates of decrease among women of either race (blacks, 0.26 percentage points per year; whites, 0.32). Trends in smoking among the various educational groups have differed markedly since 1966 (Pierce, Fiore et al. 1989b). College graduates have decreased their smoking level from 33.7 percent in 1966 to 16.3 percent in 1987. The rate of decline from 1966-85 was 0.76 percentage points per year (95percent confidence limits, 0.60 to 0.92). Smoking prevalence in respondents who reported having attended some col- lege decreased from 42.5 percent in 1966 to 26.1 percent in 1987 at a slightly lower rate of change (-0.70 percentage points per year) than that of college graduates. High school graduates who did not attend college reduced their smoking from 41.1 percent in 1966 to 33.1 percent in 1987 at a rate (-0.32 percentage points per year) less than half that for respondents who had attended college. Smoking prevalence in those respondents without a high school diploma did not change appreciably from 1966 (36.5 percent) to 1987 (35.7 percent); the rate of decline between 1966 and 1985 was only 0.06 percentage points per year. Thus, there is a twelvefold difference in rate of decline in smoking prevalence between the most and least educated groups in our society. The increasing gap in smoking prevalence by educational attainment is particularly evident when comparing the difference in smoking prevalence between the most and least edu- cated groups in 1966 with the difference in 1987. In 1966, the prevalence rates were similar (33.7 and 36.5 percent, respectively); in 1987, prevalence in the most educated group (16.3 percent) was less than half that in the least educated group (35.7 percent). Adult Use of Tobacco Surveys: 1964-86 In 1964, 1966, 1970, 1975, and 1986, the Office on Smoking and Health (formerly the National Clearinghouse for Smoking and Health) conducted detailed surveys of a representative sample of the U.S. adult population. The purpose of these surveys has been to study the population's knowledge, attitudes, and practices regarding the use of tobacco. The first two surveys primarily used in-person household interviews while the last three used telephone interviews. Prevalence of cigarette smoking in the United States as measured by the AUTSs has declined from 40.3 percent in 1964 to 26.5 per- cent in 1986 (Table 4). This decrease represents an overall decline in smoking of more than 34 percent during this 22-year period. 271 TABLE 4.Trends in smoking prevalence (o/o), AUTS versus NHIS Survey year AUTSa Estimated NHISb Difference (NHIS-AUTS) 1964 40.3 40.4 0.1 I966 42.2 39.4 -2.8 1970 36.2 31.4 1.2 1975 33.8 34.9 I.1 1986 26.5 29.4 2.9 "For all survey years. includes respondents aged 21 years and older except 1986, which includes respondents aged 17 years and older. All data weighted. %tcludes respondents aged 20 years and older. Values for each year are determined by extrapolating expected prevalence values based on regression analysis from Table 3. SOURCE: Office on Smoking and Health (US DHEW 1969.1973a. 1976; CDC 1987a). Unlike the NHIS, for which data are collected during an in-person household inter- view, AUTSs collected data via telephone interviews in 1970, 1975, and 1986. The three AUTSs conducted since 1970 all produced prevalence estimates below those es- timated (by regression analysis) from the NHISs (Table 4). The largest difference be- tween the two surveyswas2.9 percentage points in 1986. The 95-percent confidence limits around the NHIS projection for 1986 are 27.8 to 31.7 compared with limits of 25.8 to 27.3 from the 1986 AUTS; thus, the difference in estimates between the two surveys is statistically significant. A difference in sampling modalities is among the most likely explanations for this discrepancy in prevalence estimates. A similar find- ing has been noted in State-specific prevalence estimates (see below). Telephone sur- veys have a small sampling bias by excluding households lacking telephones and may have a greater nonresponse bias because of generally lower response rates compared with household surveys (CDC 1987a). Cigarette Smoking Among Different Occupational Groups NHIS data have been published on smoking prevalence by occupation for the years 1970, 1978-80 combined, and 1985 (Table 5). There is a consistent pattern of higher smoking rates among blue-collar and service workers than among white-collar workers for all these survey years. For example, in 1985, the prevalence of smoking among blue-collar and white-collar workers was 40 and 28 percent, respectively. This dif- ference was greater among males (14 percentage points) than among females (6 per- centage points). Detailed data on smoking prevalence, percentage of former smokers, quitting attempts, and age of initiation within specific occupational categories for 1978- 80 were published in the 1985 Surgeon General's Report (US DHHS 1985). Weir&am and Sterling (1987) also provided a detailed analysis of smoking by occupation using the 1970 and 1979-80 NHIS data. Novotny, Warner, and colleagues (1988) performed multivariate logistic regression analyses on data from the 1985 NHIS (ages 25 to 64 years) to examine the independent 272 TABLE S.-Prevalence of smoking (%) by occupation, 1970,1978-80, and 1985 1970' 1978-80a Occupation Males Females Males Females Males I mb Females Total Currently employed 47.9 36.S 39.9 33.3 33.8 30.0 32.1 White collar 40.8 36.1 33.0 3 I .9 26.4 28.0 27.5 Blue collar 55.0 37.7 47.1 3x. I 40. I 33.9 39.7 Service 53.3 3Y.4 47.5 37.4 40.3 35.4 37.2 Unemployed 55.9 42.3 53.1 39.6 44.3 28.0 36. I "Aged 20 to 64 years. hA&yd 20 yeam and older. SOURCE: NHlSs IY70. 197X-XO(combinedh and IY85. NCHS (US DHHS IYXS. 1988). effects of socioeconomic status (SES) and selected demographic factors on the odds of ever smoking (versus never smoking) and current smoking (versus former smoking). The SES/demographic factors included in the models were: sex, employment status, occupation, education, marital status, and poverty status. The investigators found that when they simultaneously controlled for the effects of these factors, unemployed per- sons were more likely than employed persons to be ever smokers or current smokers. However, blue-collar and service workers were not found to have significantly in- creased odds of ever or current smoking compared with white-collar workers. Employed persons were more likely to have quit smoking than unemployed persons. Special Populations: Hispanics Information on smoking among Hispanics was collected as part of the Hispanic Health and Nutrition Examination Survey (HHANES) between 1982 and 1984. This was a geographically based sample of Hispanics from three areas of the United States designed to represent three large Hispanic groups (Puerto Ricans in the New York City area; Cuban-Americans in Dade County, Florida; and Mexican-Americans inthe South- west). Sample sizes were 9,000 Mexican-Americans, 4,000 Puerto Ricans, and 1,500 Cuban-Americans. According to the HHANES, the age-adjusted smoking rates for males aged 20 to 74 years were 43 percent for Mexican-Americans, 42 percent for Cuban-Americans, and 40 percent for Puerto Ricans. Among females, the smoking prevalence was 24 percent for Mexican-Americans and Cuban Americans and 30 percent for Puerto Rican Americans (Haynes 1987). A birth-cohort analysis of these data showed that smoking rates have decreased among successive cohorts of men, but increased among succes- sive cohorts of women (Escobedo and Remington 1989). These rates are higher than those obtained from the NHISs for the years 1979 and 1980 (Marcus and Crane 1985; Rogers and Crank 1988) and 1985 (Marcus and Crane 1987). However, the number of Hispanics in these NHIS samples was small, making prevalence estimates less reliable. Haynes (1987) suggests that NHIS data may under- estimate smoking prevalence among Hispanics because questions about smoking were not asked in Spanish. The first estimates of smoking behavior among Hispanics that are both national and statistically reliable will be available from the 1987 NHIS, which oversampled for this population group. Special Populations: American Indians and Alaskan Natives There are no reliable national estimates of smoking prevalence among American In- dians. Several surveys have assessed smoking rates among specific Indian tribes or on certain Indian reservations (CDC 1987b). Smoking prevalence is highest among North- em Plains Indians (42 to 70 percent) and Alaskan Natives (56 percent), where rates greatly exceed the rate in the general U.S. population. Much lower rates have been reported for Indians from the Southwest ( 13 to 28 percent). High rates of smokeless tobacco use have also been reported among some American Indian groups, especially in Indian youth. According to a survey of approximately 5,ooO children 5 to 18 years of age in rural Alaska conducted by the Indian Health Service, 28 percent of girls and 34 percent of boys reported using smokeless tobacco products (CDC 1987~). Similar findings were obtained in other surveys of Native Americans (Schinke et al. 1987; CDC 1988; Hall and Dexter 1988). 274 Special Populations: Asian Americans There are no reliable national estimates of smoking prevalence among Asian 4mericans. A few local surveys provide estimates of smoking prevalence among Asian Americans in specific geographic regions. The State of Hawaii has a population composed of 29 percent Caucasian, 26 percent Japanese, 15 percent Hawaiian, and 15 percent Filipino. The State conducted a Be- havioral Risk Factor Survey (see below) of 1,002 people by telephone in 1984. Smok- ing prevalence estimates were 28 percent for Caucasians, 27 percent for both Hawaiians and Filipinos, and 23 percent for Japanese (Hawaii State Department of Health 1984). A similar survey of 1,557 residents of the State was completed in 1986. Prevalence es- timates from this second survey were 29.3 percent for Caucasians, 28.8 percent for Hawaiians, 25.1 percent for Filipinos, and 20.6 percent for Japanese (Chung 1986). Special Populations: Pregnant Women National data on smoking during pregnancy are scarce, especially prior to 1980. Since 1980, several national surveys have directed smoking questions to previously pregnant women, but survey methodologies vary widely and it is not possible to study secular changes in behavior. Probably the best source of national data on smoking among pregnant women has been the National Natality Surveys (NNSs). which were conducted among national samples of married mothers of live infants born in 1967 and 1980. Data from these sur- veys were used by Kleinman and Kopstein (1987) to document changes in smoking be- havior during pregnancy over that period of time. Among teenagers, smoking rates remained fairly constant over time at about 38 percent among whites and 27 percent among blacks. Among women over age 20, there were decreases in smoking prevalence that varied markedly by race and by educational attainment of the mother. Smoking prevalence among white women over age 20declined from 40 percent in 1967 to 25 percent in 1980; among black women over age 20, it declined from 33 percent to 23 percent. Among white women over age 20. there was an increase in the proportion quitting smoking during pregnancy (11 percent to 16 percent), while among blacks the proportion quitting actually decreased (17 percent to 11 percent). Among white women with less than 12 years of education, the prevalence of smoking during pregnancy declined from 48 percent to 43 percent, while for women with 16 or more years of education, it declined from 34 percent to 11 percent. Among white smokers with less than 12 years of education, there was relatively little change in the proportion quitting during pregnancy (11 percent to 9 percent), but among smokers with 16 years or more of education, the proportion more than doubled (12 percent to 27 percent). Insufficient numbers of black women were sampled to study trends by education among blacks. A study similar to the NNS, the National Maternal and Infant Health Survey, was begun in 1988. Data from that study will provide the best estimates of smoking during pregnancy for the late 1980s. At this time, however, no comparable national data exist to study women after 1980. Studies that have asked about smoking behavior during pregnancy have not asked about behavior during specific years, so it is not possible to 215