TABLE M.-Trends in smoking initiation, NHISs, United States, 1965-87 Smoking prevalence (%), ages 20-24 Education level Overall Year population 1965 47.8 1966 47.7 1970 41.5 1974 39.5 1976 39.6 1977 38.8 1978 35.4 1979 35.8 1980 36. I 1983 36.9 1985 31.8 1987" 29.5 Trend information (1965435) Changeh/year -0.69 Standard error 0.09 R2 0.86 Sex Race High school graduate or less Some college or more Males Females Whites Blacks Males Females Males Females 56.3 40.5 47.5 50.8 63.6 42.6 42.7 34.5 57.7 39.5 48.2 45.5 65.1 41.3 43.5 34.7 48.5 35.8 41.2 45.2 60.0 40.2 33.2 26.8 44.3 35.4 38.6 47.1 52.7 40.1 34.7 26.4 45.9 34.2 39.5 42.3 54. I 41.0 34.4 23.0 40.4 37.4 38.5 41.5 52.2 43.0 24.0 27.5 38.5 32.5 35.7 34.8 46.8 39.3 25.9 21.1 37.7 34.0 35.6 36.7 47. I 4 I .9 23.8 22. I 40.0 32.5 35.9 37.9 50. I 40.3 20.1 19.4 36.9 37.0 36.8 38.7 49.1 45.5 16.2 22.9 31.0 32.5 32.5 28.2 43.0 43.6 15.5 17.2 31.1 28.1 30.5 25.6 43.8 37.6 16.3 15.1 -1.19 -0.28 4.68 a.79 -1.00 0.10 -1.51 -0.72 0.10 0.12 0.09 0.17 0.13 0.10 0.13 0.15 0.94 0.40 0.85 0.7 I 0.87 NA' 0.95 0.75 aProvisional data only. s % percentage Points. `The slope of the regression line was not significantly different from zero, making the R2computation inappropriate. SOURCE: NH& 1965-87: unpublished data, Office on Smoking and Health. only from 40.5 percent in 1965 to 28.1 percent in 1987 at a rate of change (1965-85) one-quarter that of young males (-0.28 percentage points per year). The slower rate of decline among women is due, in large part, to the increase in initiation rates in less edu- cated young women (Pierce, Fiore et al. 1989 b). Smoking initiation patterns among whites and blacks have been similar during the past 20 years. From 1965-87, smoking prevalence among whites aged 20 to 24 years has decreased from 47.5 percent to 30.5 percent, while for blacks the decline has been from 50.8 percent to 25.6 percent. The rates of change between 1965 and 1985 among whites and blacks were similar (-0.68 and -0.79 percentage points per year, respective- ly). The prevalence of smoking had been higher among young blacks than among young whites for most survey years between 1965 and 1983, but whites had a higher prevalence in 1985 and 1987. Marked differences in smoking initiation rates based on educational level have oc- cut-red. From 1965-87, the smoking initiation rate as measured by prevalence, ages 20 to 24, fell among males with 12 or fewer years of schooling (high school graduate or less) from 63.6 percent to 43.8 percent (-1 .OO percentage point per year from 1965- 85). In contrast, for males with 13 or more years of schooling (some college or more), prevalence has fallen from 42.7 percent to 16.3 percent, at a rate of decline (1965-85) of 1.5 1 percentage points per year. A similar difference in initiation rates by education was seen among women, although the rate of decline between 1965 and 1985 was less among women than among men of equivalent education. In the overall sample (men and women combined), the rate of decrease in initiation among persons with 13 or more years of education (1.10 percentage points per year) was three times that among per- sons with 12 or fewer years of education (0.35). Trends in Adolescent Smoking Several surveys have provided national estimates of smoking prevalence among adolescents. Because these surveys differ in terms of the definitions of smoking, ages of respondents, sample size, method of data collection (household versus school ver- sus telephone interview), years in which the surveys were conducted, and overall results, the findings of the major surveys are presented below. NIDA High School Seniors Surveys on Drug Use, 197687 Data from the NIDA-sponsored High School Seniors Surveys have been collected annually since 1975 and are presented in Table 19. These surveys have been carried out by the University of Michigan Institute for Social Research (Johnston, O'Malley, Bachman 1987). This data set is most useful for examining trends in smoking. In- dividual prevalence figures probably underestimate actual adolescent smoking prevalence because the survey does not include high school dropouts, who are known to have much higher smoking rates (Pirie et al. 1988; Yates et al. 1988). Reported daily smoking of cigarettes has decreased among high school seniors from a peak prevalence of 29 percent in 1976 to 19 percent in 1987. However, the trend has not been linear. The majority of the change occurred between 1978 and 1980, after 302 TABLE 19.-Smoking status (%) of high school seniors, United States, 197M87 Y.S Daily Less than smokers daily smokers Previous smokers, not in last month Never smokers 1975 1976 1977 1978 1979 1981 1982 1983 1984 1985 1986 1987 27 10 37 26 29 IO 36 25 29 10 38 24 28 9 38 25 26 9 40 26 21 9 41 29 20 9 42 29 21 9 40 30 20 9 41 29 18 II 41 30 19 11 39 31 18 II 38 32 19 II 38 33 SOURCE: Institute forSocial Research,Universityof Michigan(Bachman.Johnston. O'Malley 19ROa.b. 198 I, 1984, ,985, 1987; Johnston and Bachman 1980; Johnston, Bachman. O'Malley 1980&b. 1982. 1984,1986. and unpublished data, 1987). which prevalence has remained relatively stable. The proportion of high school seniors who have smoked within the last month, although not on a daily basis, has not changed substantially during this period. There is also rather little change in the proportion of this population who has previously smoked but not in the last 30 days. The proportion of high school seniors who have never smoked increased from 26 percent to 33 percent between 1975 and 1987. Trends in smoking status by sex, race, and educational plans are presented in Table 20. The prevalence of daily smoking decreased in all major subcategories of high school seniors between 1976 and 1987. Daily smoking among males decreased from a peak prevalence of 28 percent in 1976 to 16 percent in 1987; most of this drop oc- curred between 1977 and 1980. Daily smoking among females decreased from a peak prevalence of 30 percent in 1977 to 20 percent in 1987, with the largest decrease oc- curring from 1979-8 1. Since 198 1, the prevalence of daily smoking among high school students has remained fairly constant for both males and females. In each year since 1977, the prevalence of daily smoking has been higher in females than in males (median difference=4 percentage points). The prevalence of daily smoking fell substantially among blacks,from 26 percent in 1976 to 8 percent in 1987. During the same period, prevalence declined among whites from 29 percent to 20 percent. The reasons for the dramatic decline among blacks are unclear. It does not appear to be due to increasing sampling bias over time-survey methods and sample sizes by race have been consistent. A substantial decrease in smok- ing initiation among blacks also occurred, as measured in the NHIS by prevalence in persons 20 to 24 years of age, between 1983 (38.7 percent) and 1985 (28.2 percent) (Table 18). This figure declined further to a preliminary estimate of 25.6 percent in 1987. Students with plans to pursue higher education were much less likely to be daily smokers in 1976 than those without such plans (21 percent versus 37 percent). The ab- 303 TABLE 20.4moking status (%) of high school seniors by sex, race, and educational plans, United States, 1975-87 Year Daily smokers Less than daily smokers Sex Race Plans for higher education Sex Race Plans for higher education M F w B Yes No M F W 6 Yes No I975 1976 1977 1978 1979 1980 1981 1982 I983 1984 I985 1986 27 26 28 28 28 30 26 29 22 2x I8 24 18 22 18 24 19 23 16 21 17 21 I7 20 16 20 29 26 21 37 28 25 20 38 27 22 18 36 26 19 17 35 22 16 I4 31 20 I3 I3 30 23 I2 I3 30 22 12 14 30 20 8 II 29 20 II 13 31 21 8 I2 29 20 8 I4 30 IO 10 10 10 10 10 9 10 9 9 8 IO 8 IO 9 9 9 IO IO II IO II II II II II 10 I3 IO 10 9 II 10 9 9 9 9 9 9 9 9 9 9 IO 9 IO 9 9 9 9 9 9 9 9 9 9 10 9 II 9 II II II 8 IO II 12 7 11 IO I2 6 II II TABLE 20.-Continued Previous smokers, not in last month Never smokers Sex Race Plans for higher education Sex Race Plans for higher education Year M F W B Yes No M F W B Yes No 1975 38 36 24 28 I976 38 36 37 36 39 35 24 2s 2s 24 31 19 1977 39 35 37 49 41 3.5 24 25 2s 26 30 19 1978 40 38 38 40 42 35 26 24 2s 29 31 20 1979 42 38 39 41 42 37 27 2s 26 30 32 20 1980 43 39 40 45 44 37 30 28 29 30 34 23 1981 43 41 41 45 45 38 31 27 29 33 33 24 1982 41 39 40 43 43 37 32 28 29 36 35 24 1983 41 40 40 45 43 38 31 28 29 34 34 24 1984 41 39 40 42 42 38 33 29 29 40 35 24 1985 39 39 38 42 41 36 33 30 30 39 36 24 1986 38 38 38 41 39 37 34 31 30 44 37 25 1987 38 38 38 41 39 35 35 31 30 45 37 25 SOURCE: Institute for Social Research. University of Michigan (See Table 19 for citations). solute difference (in percentage points) between the two groups remained constant be- tween 1976 and 1987. In 1987, the prevalence of daily smokers among those with plans for higher education was less than half the prevalence among those without such plans (14 percent versus 30 percent). The percentage of blacks who smoke on less than a daily basis exceeded the percent- age of whites in 1976 (13 and 10 percent, respectively) but was lower than the percent- age of whites in 1987 (6 and 12 percent, respectively). The percentage who have pre- viously smoked but not in the past month has consistently been slightly higher among blacks than among whites and among those with plans for higher education than among those without college plans. Besides these findings, there have been few differences between subgroups and few changes between 1976 and 1987 in the proportion of high school seniors who are in these categories. As mentioned above, the decrease in the proportion of high school seniors who smoke on adaily basis is reflected by a complementary increase in the proportion of high school seniors who have never smoked. This increase has been more marked among males compared with females and among blacks compared with whites. 1987 National Adolescent Student Health Survey The 1987 NASHS collected data on prevalence of smoking within the last 30 days (US DHHS, in pressb). Respondents to this survey composed a random sample of the Nation's students in 8th and 10th grades. Sixty-three percent of the 8th graders were 13 years old and 27 percent were 14 years old. Sixty-two percent of the 10th graders were 15 years old and 28 percent were 16 years old. For each grade, 68 percent were white, 17 percent were black, and 9 percent were Hispanic. Prevalence data are presented in Table 2 1. Eighty-four percent of the eighth graders reported that they had not even puffed on a cigarette in the last 30 days, with little dif- ference between the sexes. Forty-nine percent of all eighth graders reported never having smoked a cigarette, with no difference between the sexes. Among 10th graders, the proportion not having puffed on a cigarette in the last 30 days was slightly lower: 76 percent among males and 7 1 percent among females. Thirty-eight percent of males and 36 percent of females in this grade reported that they had never had a cigarette. TABLE 21.-30-day prevalence of smoking (%), United States, 1987,8th and 10th grades 8th grade 10th grade Males Females Males Females Not even a puff 84.9 83.0 75.9 71.3 14 cigarettes 7.1 8.2 7.8 10.4 5-19 cigarettes 2.7 3.4 4.8 5.1 l-5 packs 2.4 3.5 5.6 7.4 More than 5 packs 2.9 I .9 6.0 5.8 SOURCE: National Adolescent Student Health Survey 1987 (US DHHS, in press, b). 306 Approximately equal proportions (7 to 8 percent) of males and females in the eighth grade reported smoking a pack or more in the last month. Among 10th graders, this proportion was more than twice as high, with 17 percent of males and 19 percent of females reporting that they smoked a pack or more in the last month. US DHEW Teenage Smoking Surveys, 1968-79 Detailed questions on smoking were asked in five national telephone surveys of adolescents (ages 12 to 18 years) conducted by Chilton Research Services for the U.S. Department of Health, Education, and Welfare from 1968 through 1979 (US DHEW 1979b). Adolescents were classified by smoking status as follows: never smokers, had not taken even a few puffs of a cigarette; experimental smokers, had had a few puffs but had not smoked as many as 100 cigarettes; es-smokers, had smoked at least 100 cigarettes but no longer smoked; current occasional smokers, smoked less than one cigarette per week; and currentregularsmokers, smoked at least one cigarette per week. In published results for these surveys, data for never smokers and experimental smokers were generally aggregated. Summary data from each of the surveys are presented in Table 22 (males) and Table 23 (females). The proportion of both males and females of each age group who are classified as either never smokers or experimental smokers is substantially higher than the proportion of never smokers reported by other surveys. For example, the 1979 Teenage Smoking Survey showed that 75 percent of males and 82 percent of females aged 15 to 16 years had never smoked or had only experimented with cigarettes; in con- trast, the 1987 NASHS (above) showed that only 38 percent of males and 36 percent of females in the 10th grade (15 to 16 years old) had never had a cigarette. Similarly, the 1979 Teenage Smoking Survey showed that 68 percent of males and 64 percent of females aged 17 to 18 years were either never smokers or experimental smokers; in contrast, the 1979 High School Seniors Survey showed that 27 percent of males and 25 percent of females were never smokers. There are at least two possible explanations for the consistently and surprisingly high proportion of teenagers in the categories of never smokers and experimental smokers. First, 100 cigarettes may be too high a cutoff to use for classifying teenagers as never smokers or experimenters. Second, telephone interviewing may lead to more under- reporting of cigarette smoking behavior than other survey modalities. Underreporting may be more important for some smoking categories than others-for instance, oc- casional smokers might be particularly sensitive about their smokjng behavior and might be more likely to underreport the total number of cigarettes they have ever smoked. Current smoking rates can also be compared between the Teenage Smoking Surveys and the High School Seniors Surveys. In the 1979 telephone survey, teenagers were classified on their reported smoking on a weekly basis. Of males aged 17 to 18 years, 19.3 percent were classified as current regular smokers (one or more cigarettes per week) and another 0.3 percent were classified as current occasional smokers (less than one cigarette per week). For females aged 17 to 18 years, these figures were 26.2 per- cent and 0.8 percent, respectively. In the High School Seniors Survey, students are 307 TABLE ZZ.-Cigarette smoking among teenage males, United States, 1968-79 Age 12-l 4 years 15- I6 years 17-l 8 years Total Smoking status Year N %' N % N % N 92 Never smoked or experimented only Former smoker Current occasional smoker 196X 876 93. I 465 75.2 1970 512 90.5 268 70.5 1972 533 91.1 273 68.3 I974 496 90.7 253 69.5 I979 527 92.X 284 15.3 196X 25 2.7 34 5.5 1970 21 3.7 35 9.2 1972 20 3.4 50 12.5 1974 28 5.1 45 12.4 1979 23 4.0 38 10.1 1968 13 1.4 14 2.3 1970 1 0.2 3 0.8 1972 5 0.9 6 1.5 1974 0 0.0 0 0.0 1919 0 0.0 4 1.1 344 54.1 1,685 71.0 178 48.1 958 72.8 211 54.4 1,017 74. I 202 55.3 951 74.5 254 68.1 1,065 80.8 71 11.3 130 5.9 52 14.1 108 8.2 56 14.4 126 9.2 44 12.1 117 9.2 46 12.3 107 8. I 24 3.8 51 2.3 2 0.5 6 0.5 4 1 .o 15 1.1 6 1.6 6 0.5 1 0.3 5 0.4 TABLE 22.-Continued Smoking status Current regular smoker Year 1968 1970 1972 1974 12-14 years IS-16 years I7- I8 years Total N % N % N % N % 27 2.9 105 17.0 190 30.2 322 14.7 32 5.7 74 19.5 138 37.3 244 18.5 27 4.6 71 17.8 117 30.2 215 15.7 23 4.2 66 18.1 113 31.0 202 15.8 Total 1979 IX 3.2 51 13.5 72 19.3 141 10.7 1968 941 100 618 100 629 IO0 2,188 100 1970 566 100 380 100 370 loo 1,316 100 1972 585 100 400 1974 547 100 364 1979 568 100 377 SOURCE: US DHEW Teenage Smoking Surveys. 1968.1970,1972,1974, 1979 (US DHEW 1979b). 100 388 100 1,373 100 100 365 100 1,276 100 100 373 100 I.318 100 TABLE 23.-Cigarette smoking among teenage females, United States, 1968-79 Smoking status Year Age I2- I4 years IS-16 years 17-18 years Total N 96 N % N 96 N % Never smoked or experimented only Former smoker Current occasional smoker I968 919 97.9 552 84.4 462 1970 536 95.0 312 81.5 264 1972 569 95.3 312 77.0 277 1974 495 90.2 250 69.3 228 1979 514 92.3 319 81.8 239 1968 7 0.7 25 3.8 38 1970 8 1.4 15 3.9 22 1972 II I .8 26 6.4 30 1974 26 4.7 33 9.1 42 1979 19 3.4 23 5.9 34 1968 7 0.7 14 2.1 15 1970 3 0.5 1 0.3 5 1972 0 0.0 1 0.2 3 1974 I 0.2 5 1.4 2 1979 0 0.0 2 0.5 3 73.0 70.0 66.7 62.1 63.9 6.0 5.8 7.2 II.4 9.1 2.4 1.3 0.7 0.5 0.8 1,933 86.8 1,112 84.0 1,158 81.7 973 76.2 I ,072 81.2 70 3.1 45 3.4 67 4.7 101 7.9 76 5.8 36 1.6 9 0.7 4 0.3 8 0.6 5 0.4 TABLE 23.-Continued Smoking status Year Age 12-14 years 15-16 years 17-18 years Total N % N % N 8 N % Current regular 1968 6 smoker 1970 17 1972 17 1974 27 1979 24 Total 1968 939 1970 564 1972 597 1974 549 0.6 63 3.0 55 2.8 66 4.9 73 4.3 46 loo 654 loo 383 loo 405 100 361 9.6 14.4 16.3 20.2 Il.8 loo loo 100 loo 118 18.6 187 8.4 86 22.8 158 11.9 105 25.3 188 13.3 95 25.9 195 15.3 98 26.2 168 12.7 633 100 2,226 100 377 100 1.324 loo 415 loo 1,417 100 367 loo 1,277 100 1979 557 100 390 100 374 100 1,321 loo SOURCE: US DHEW Teenage Smoking Surveys, 1968.1970,1972,1974,1979 (US DHEW 1979b) classified based on their reported smoking during the past 30 days. In the 1979 High School Seniors Survey (Table 22). 22 percent of males were classified as daily smokers and another 9 percent reported having smoked in the last month but not on a daily basis. In the same year, 29 percent of females were daily smokers and 9 percent smoked on less than a daily basis. Comparing these two data sets shows that the telephone survey obtained lower es- timates for weekly smoking than the school survey obtained for daily smoking ( 19 vs. 22 percent for males, 26 vs. 28 percent for females). The remaining current smokers (defined as less than one cigarette per week in the telephone survey and less than one per day in the school survey) were also estimated at lower rates in the telephone sur- vey (0.3 vs. 9 percent for males, 0.8 vs. 9 percent for females). This suggests that the telephone survey underestimated both the number of daily smokers and the number of less-than-daily smokers. Most of the discrepancy appears to be due to a failure to iden- tify the latter. It is unclear whether this difference is related to the system of classifying smokers or the telephone survey methodology. NIDA National Household Surveys on Drug Abuse, 1979-85 NIDA conducted household surveys on drug abuse in 1979, 1982, and 1985. For each of these surveys, data were obtained from a stratified random sample of 8,ooO U.S. households; approximately 2,000 in-person interviews were conducted with respon- dents 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. These surveys indi- cated that approximately 26 percent of the teenage population surveyed smoked at least one cigarette at some time during 1985 (Table 24). In 1985,15.6 percent of this popula- tion had smoked within the previous month. Comparisons between data from the 1979 household survey and data from the more recent surveys are not appropriate, because in 1979 prevalence of use within the past year or past month was reported only for those who had smoked 100 cigarettes in their lifetime; this lifetime cutoff was not used in the later surveys. TABLE 24.-Prevalence (so) of cigarette use among youth 12 to 17 years of age, 1979,1982, and 1985, United States Survey year Any use in Used in last last year 3odays 1979a 13.3 12.1 1982 24.8 14.7 1985 26.0 15.6 `The 1979 estm~ates are not necessarily comparable to later estimates because the 1979 survey asked questions only of those who had smoked 100 cigarettes in their lifetime. SOURCE: NIDA National Household Surveys on Drug Abuse 1979. 1982. 1985 (US DHHS 1988). 312 Summary Several national surveys provide information on adolescent smoking. These surveys vary substantially in sample size, methodology, definitions of smoking, ages of respon- dents, and other factors that may appreciably affect prevalence estimates. The best trend data are available from the annual high school seniors survey. This survey shows that prevalence of daily cigarette consumption declined from 29 percent of seniors in 1976 to 21 percent in 1980, after which prevalence leveled off at 18 to 21 percent. Smoking prevalence among females has consistently exceeded that among males since 1977. The leveling off of smoking prevalence among high school seniors raises concern that the steadily declining initiation rates as determined by prevalence among adults aged 20 to 24 (NHIS) may soon level off as well. Smoking prevalence has been consistently lower for high school seniors with plans to pursue higher education than for those without such plans. In 1987, smoking rates were 14 and 30 percent in these two groups, respectively. Differences in prevalence of smoking and smokeless tobacco use (see below) be- tween young males and young females suggest that the prevalence of any tobacco use is similar in these two groups. Whereas the prevalence of smoking is higher among female high school seniors than among males, the prevalence of smokeless tobacco use is higher among young males than among young females. Changes in the Types of Cigarettes Smoked Data on the market share of filter and nonfilter cigarettes, cigarettes of different machine-determined "tar" and nicotine yields, menthol and nonmenthol cigarettes, and cigarettes of different length have been published by the Federal Trade Commission (FTC) from information supplied to the agency by the major cigarette companies. Filtered Cigarettes Filters are the design characteristic of commercial cigarettes that most affects their machine-measured yield of harmful constituents (US DHHS 1981). Filters selectively remove nitrosamines and semivolatile phenols from smoke. Thus, filters affect not only the absolute amounts of these constituents delivered in smoke but also their relative concentrations in cigarette "tar. " Since the early 195Os, the proportion of cigarettes in the United States sold as filtered cigarettes has increased steadily. In 1950, less than 1 percent of cigarettes sold in the United States were filtered. That proportion rose to 19 percent in 1955,51 percent in 1960, and 94 percent in 1986 (Table 25). Low-Tar, Low-Nicotine Cigarettes Trends in the sales-weighted average yield of tar and nicotine for cigarettes sold in the United States are shown in Figure 14 of Chapter 2. The sales-weighted average is based on the tar and nicotine yield of specific brands (as measured by the FTC machine- 313 TABLE 25.Domestic market share of filter cigarettes as a proportion of total cigarettes sold United States, 1950436 Year Market share (%) YCU - Market share (%) - 1950 0.6 1969 77 1951 0.7 1970 80 1952 1 1971 82 1953 3 1972 84 1954 9 1973 85 1955 19 1974 86 1956 28 1975 87 1957 38 1976 88 1958 45 1977 90 1959 49 1978 90 1960 51 1979 91 I%1 52 1980 92 1962 55 1981 92 I963 58 1982 93 1964 61 1983 93 196.5 64 1984 93 1966 68 1985 94 1967 72 1986 94 1968 74 SOURCE: FTC (1988). testing method) multiplied by the quantity of sales for those brands. The sales-weighted average yield of tar fell from 35 mg in 1957 to 13 mg in 1987. For nicotine, the sales- weighted average fell from 1.3 mg in 1968 to 1.0 mg in 1985. However, the sales- weighted average yield of tar and nicotine leveled off between 1981 and 1987. As pointed out in Chapter 2, modifications in the makeup of commercial cigarettes have profoundly influenced these yields; for example, the steepest declines occurred in the late 1950s after introduction of filter tips. Trends in the percentage of domestic sales of cigarettes yielding lower tar levels are shown in Table 26. The domestic market share of cigarettes yielding 15 mg or less tar increased from 2.0 percent in 1967 to 56.0 percent in 1981. Since 1981, this propor- tion has fallen slightly and has stabilized at 51 to 53 percent. About two-thirds of these cigarettes have tar yields between 9 and 15 mg. It should be noted that the parameters used in the FTC machine-testing method (developed in the 1960s) do not necessarily reflect current smoking patterns. For ex- ample, the FTC method uses one puff per minute (Pillsbury et al. 1969), whereas human 314 TABLE 26.--Domestic market share of cigarettes with reduced tar, percentage of total cigarettes sold, United States, 1967436 Tar Yield YeiS 215 mg 512 mg S9 mg 16 mg <3 mg 1967 2.0 1968 2.5 1969 3.0 1970 3.6 1971 3.8 1972 6.6 1973 8.9 1974 8.9 1975 13.5 1976 15.9 1977 22.7 1978 27.5 1979 40.9 10.6 5.8 2.7 1980 44.8 16.8 7.3 3.3 1981 56.0 24.6 9.6 3.7 1982 52.2 43.8 27.8 8.9 2.9 1983 53.1 44.9 27.9 9.4 3.1 1984 51.0 43.4 26.3 9.4 2.9 1985 51.9 43.1 25.3 8.4 2.3 1986 52.6 44.5 22.3 9.9 2.6 SOURCE: FI-C (1988); Kozknwki (1989). studies of smoking patterns show an average interpuff interval of 34 seconds (that is, about two puffs per minute) (US DHHS 1988, Chapter 4, Table 2). According to the 1986 AUTS, 41 percent of smokers smoke cigarettes yielding 15 mg or less tar (Table 27). The proportion of smokers smoking cigarettes yielding more than 15 mg tar is higher among males, blacks, and persons with less`education com- pared with females, whites, and more educated persons, respectively. This proportion decreases with age; the higher proportion among those 17 to 19 years of age probably reflects the popularity of the higher tar Marlboro brand among adolescents (Hunter et al. 1986; Goldstein et al. 1987; Glantz 1985). Increased consumer demand for lower yield cigarettes during the past two decades is probably attributed to consumer beliefs that lower yield brands are less hazardous. This impression may have resulted in part from cigarette advertising implying that IOW- yield brands are less hazardous or are safe (Davis 1987). According to the 1986 AUTS, 45 percent of current smokers believe that some kinds of cigarettes are probably more hazardous than others (see Chapter 4). 315 TABLE 27.-Percentage of current smokers, aged 17 years and older, who use cigarettes of varying tar yields and who use menthol cigarettes, by sex, race, and education, 1986 Percentage of current smokers 510 Tar yield (mg/cigarette) >I&15 >I5 Total Menthol cigarette smokers Total Sex Males Females Age 17-19 2624 25-44 454 265 Race White Black Other Education