Parental Influences on Adolescent Marijuana Use and the Baby Boom Generation

CHAPTER 2: EXPOSURE TO THE MARIJUANA EPIDEMIC

2.1 A Basic Hypothesis

A basic hypothesis of the study was that child marijuana use would be most strongly associated with parental marijuana use among families where the parent was a member of baby boom generation cohorts that were exposed to the marijuana epidemic in adolescence. This hypothesis was based on the assumption that cultural and historical factors experienced by parents influence involvement in marijuana use by their children. This would be reflected in a greater association in marijuana use within parent-child dyads among baby boom families exposed to the marijuana epidemic than among those not exposed or in non-baby boom families. Cohorts born between 1946 and 1964 constitute the baby boom generation (Light, 1988).

An empirical test of the hypothesis required a clear definition and delineation of the marijuana epidemic, the historical period when it was manifest, and parental exposure to the epidemic in critical periods of development, when risk for marijuana initiation is the highest.

2.2 Definition of Exposure to the Marijuana Use Epidemic

In classical epidemiology, the beginning of an epidemic is usually conceptualized in terms of incidence rates, with the first case marking the beginning of the epidemic and exposure to it adjusted for any relevant incubation period (Kelsey, 1996). Such a criterion may not be appropriate to a drug epidemic, where modeling and socialization play important roles and may depend upon a critical mass of users. The concept of an epidemic may not even be appropriately applied to historical changes in marijuana use. While traditional epidemics may have an insidious onset, the experience of a marijuana epidemic requires an explicit awareness of patterns of behavior in society.

We considered that two features of marijuana consumption in the population may affect individuals' perceptions and experiences of cultural changes related to marijuana use: incidence, i.e., the rate of new users in the population, and prevalence, i.e., the rate of users at any particular time. We conceptualized that the adolescent experience of the marijuana epidemic involves exposure to two features of marijuana use in the population: exposure to high (or increasing) incidence rates and exposure to high prevalence rates of use, especially among young adults. High prevalence results from high incidence rates and sustained use after increases in incidence. Different parental birth cohorts were exposed to different rates of incidence and prevalence of use in their adolescence. The differentiation between incidence and prevalence constitutes a novel way of conceptualizing exposure to the marijuana epidemic.

2.3 Identification Of Historical Incidence and Prevalence Periods

We examined changes in yearly rates of incidence and prevalence of marijuana use in the population to identify historical periods that marked the marijuana use epidemic and to delineate which birth cohorts experienced the epidemic in adolescence, according to the incidence and prevalence criteria. A problem in using these concepts to define the experiences of various birth cohorts is that incidence and prevalence rates do not show sharp changes, especially in the declining phase, so that the delineation of boundaries for the epidemic is somewhat imprecise. We examined data from the NHSDA and Monitoring the Future Study to identify these historical periods. As regards incidence, we extended the analyses conducted by Gfroerer and Brodsky (1992), Johnson and Gerstein (1998) and Johnson et al. (1996) in the NHSDA of the number of initiates by increasing the number of survey years both backward and forward in time and disaggregating the five-year grouped birth cohorts into single years. Because of differences in the age structure of the U.S. population over time, we also estimated incidence rates for onset for all ages, and by ages 16 and 19, by dividing the number of initiates by the size of the age-specific population at risk for initiation. [See Johnson and Gerstein (1998) for a discussion of the difficulties involved in specifying a denominator.] The incidence analysis generated approximate numbers because of imprecision in the denominator used to calculate the rates. The census-based population data for each year of historical interest, years 1962 to 1996, were only available for five-year age groups. In addition, we did not subtract from the base population sample the number of youths who had already started to use marijuana by each age up to age 19 and were no longer at risk for onset. We assumed that while this might bias somewhat the absolute rates it would not affect significantly the shape of the curve over the thirty-year interval. This assumption may be incorrect, however. We used as a denominator the number of persons aged 10-19 in each of the survey years. While the boundaries are somewhat imprecise, they are probably valid within a couple of years.

As regards prevalence, in addition to specific distributions of last year marijuana use in published NHSDA reports for 14 surveys (years 1974, 1976, 1977, 1979, 1982, 1985, 1988 and 1990-1996), we also examined last year use reported by high school seniors in Monitoring the Future for years 1976-1997 (Johnston et al., 1998; U Michigan, 1998).

Figure 2.1 shows the estimated number of marijuana use initiates in the total population from Gfroerer and Brodsky (1992) based on NHSDA 1985-1991 and a replication based on NHSDA 1979-1996. The number of marijuana use initiates in the total population increased from 1962 to 1972, peaked in the years 1973 to 1977, and beginning in 1978 declined gradually through 1989. The peak years were the same for individuals who started using marijuana by age 16 and those who started later by age 19 (Figure 2.2).

The same trend curve is observed for percentages of marijuana initiates (Figure 2.3) as was observed for number of initiates.

Since the NHSDA surveys were not administered annually until 1990, year-specific prevalence rates for prior years cannot be estimated (SAMHSA, 1998b). Table 2.1 displays trends in the prevalence of lifetime and last year use in the total population, while Figure 2.4 displays last year rates of marijuana use separately for age groups 12-17, 18-25 and 26 and older or 26-34, depending on the availability of published data for different survey years. The highest rates of last year use by those 26 years or older peaked in 1981-1984, while the rates for younger respondents peaked two years earlier in 1979. Similarly data from Monitoring the Future indicate that the rates of lifetime marijuana use by 12^th graders peaked in 1979-1980 and rates of last year use in 1979 (Figure 2.5).

To highlight similarities and differences in incidence and prevalence patterns, Figure 2.6 displays on the same graph incidence rates, i.e., the estimated percent of initiates by age 19 from 1962 to 1994, as well as prevalence rates, i.e., the percent of last year marijuana use among 18 to 25 year olds from 1974 to 1996 in the NHSDA. While incidence rates by age 19 peaked in 1977, the highest marijuana prevalence in the NHSDA occurred in 1979 for adolescents and young adults.

Based on changes in rates of incidence and last year prevalence identified from these analyses and from published findings, described above, we delineated five historical periods that defined different periods of the marijuana epidemic and characterized the cultural context with respect to marijuana use of the parents in their adolescence.

1. 1963 and earlier: Pre epidemic; low incidence and prevalence of marijuana use.

2. 1964-1971: Low marijuana incidence; the number of new users began to increase.

3. 1972-1977: High marijuana incidence; the number of new marijuana users peaked and remained high among all age groups.

4. 1978-1982: High marijuana use prevalence; the number of new users gradually decreased but the rates of prevalence were at their highest levels, especially among young adults 18-25 years old.

5. 1983 and later: Post-epidemics; incidence and prevalence rates declined.

2.4 Identification of Ages at Risk for Initiation of Marijuana Use:

To place the relevant adolescent experience of different birth cohorts in the context of the five historical phases of the marijuana epidemic, it was necessary to define the developmental periods of greatest risk for initiation to marijuana use. Although the average age of marijuana onset decreased slightly over time from 1964 to 1989 (Gfroerer and Brodsky, 1992), for simplicity all cohorts were considered together.

To delineate the ages at highest risk for initiation of marijuana use, we examined the distribution of ages of onset among the self-reported users in the 1979-1996 NHSDA surveys (Table 2.2). We also calculated the hazard rate of marijuana use initiation (Figure 2.7). Since the hazard rates take right censoring (i.e., initiation may occur at a later age for many young respondents in the NHSDA) into account, they provide more precise estimates of the age-related risk for marijuana onset. Both the age-specific distribution of initiates and the hazards converged in highlighting ages 15-18 as the years of highest risk. Kandel and Logan (1984) had earlier shown by following a cohort over time that the hazard for onset of marijuana use starts to increase at age 13 and peaks at age 18. We concluded that ages 15-18 constituted the adolescent years of highest risk for marijuana initiation. These ages were used to characterize the duration of exposure to historical periods of marijuana use (or marijuana epidemic) for each cohort.

Appendix Figure A.2.1 illustrates the slight increases in the steepness of the hazard curves and shifting to the left among cohorts born between 1940 and 1984. This confirms Gfroerer and Brodsky's (1992) conclusion noted above that the average age of marijuana use onset decreased slightly from 1964 to 1989.

2.5 Cohort-Specific Exposure to Different Periods of the Epidemic

In a next step, we identified the parental birth cohorts who experienced different periods of the marijuana epidemic during ages 15-18. Some birth cohorts spent the years 15-18 entirely in one historical period ("pure" exposure); other birth cohorts spent the years 15-18 in two adjacent periods ("mixed" exposure) (Table 2.3).

Nine types of cohorts were identified as listed on Table 2.4.

The delineation of these nine groups of cohorts provided the basis for examining the impact of parental membership in the baby boom generation on child marijuana use. Since very few parents fell into the two post epidemic cohorts, these two groups were aggregated in the analysis.

Table 2.1. Trends in Prevalence of Lifetime and Last Year Marijuana Use by Age¹ (NHSDA 1974-1996)

	1974	1976	1977	1979	1982	1985	1988	1990	1991	1992	1993	1994	1995	1996
	%	%	%	%	%	%	%	%	%	%	%	%	%	%
Lifetime
12-17 years	23.0	22.4	28.0	26.7	23.2	20.1	15.0	12.7	11.1	9.1	9.9	13.6	16.2	16.8
18-25 years	52.7	52.9	59.9	66.1	61.3	57.6	54.6	50.4	48.8	46.6	45.7	41.9	41.4	44.0
26-34 years	-	-	-	45.0	51.5	54.1	57.6	56.5	55.2	54.3	54.9	52.7	51.8	50.5
26 + years	9.9	12.9	15.3	-	-	-	-	-	-	-	-	-	-	-
35 + years	-	-	-	9.0	10.4	13.9	17.6	19.6	21.1	22.2	23.8	25.4	25.3	27.0
Last Year
12-17 years	18.5	18.4	22.3	21.3	17.7	16.7	10.7	9.6	8.5	6.9	8.5	11.4	14.2	13.0
18-25 years	34.2	35.0	38.7	44.2	37.4	34.0	26.1	23.0	22.9	21.2	21.4	21.4	21.8	23.8
26-34 years	-	-	-	20.5	21.4	20.2	14.2	14.4	11.6	11.5	11.1	11.5	11.8	11.3
26+ years	3.8	5.4	6.4	-	-	-	-	-	-	-	-	-	-	-
35 + years	-	-	-	4.3	6.2	4.3	3.7	4.2	4.6	3.8	4.6	4.1	3.4	3.8

¹Adjusted estimates prepared by SAMHSA for year 1979-1996 for four age groups. Data for 1974-1977 aggregate all respondents aged 26 and older.
Sources: SAMHSA (1991; 1998b).

Table 2.2. Distribution of Self-Reported Ages of Onset into Marijuana
Use¹ Among Users in Aggregate NHSDA 1979-1996 Surveys

		1979-1996
Age of Onset		N			%
Before 10		799			0.9
10		610			0.7
11		991			1.1
12		2,976			3.4
13		4,682			5.5
14		6,304			7.8
15		7,945			10.2
16		10,471			14.3
17		7,384			11.1
18		6,942			11.4
19		3,512			6.3
20		2,790			5.4
21		2,042			4.3
22		1,145			2.7
23		758			1.9
24		539			1.2
25		802			2.0
After 25		2,380			9.7
Total N		63,072			100%

¹Weighted estimates, unweighted N's.
Sources: SAMHSA, Office of Applied Studies, National Household Survey on Drug Abuse

Table 2.4. Birth Cohorts by Type of Exposure to the Marijuana Epidemic

Birth Cohort	Type of Exposure	Historical Pattern of Marijuana Use	Total N
(1) 1945 and before (2) 1946-48* (3) 1949-53* (4) 1954-56* (5) 1957-59* (6) 1960-62* (7) 1963-64* (8) 1965-67 (9) 1968 and after	(pure) (mixed) (pure) (mixed) (pure) (mixed) (pure) (mixed) (pure)	Pre-epidemic Pre-epidemic/low incidence Low incidence Low incidence/high prevalence High incidence High incidence/high prevalence High prevalence High prevalence/post- post epidemic Post-epidemic	2,119 1,066 1,951 1,235 1,379 1,165 366 166 16

* Members of baby boom generation

This page was last updated on June 03, 2008.

SAMHSA, an agency in the Department of Health and Human Services, is the Federal Government's lead agency for improving the quality and availability of substance abuse prevention, addiction treatment, and mental health services in the United States.