Parental Influences on Adolescent Marijuana Use and the Baby Boom Generation

CHAPTER 5: PREVALENCE AND ASSOCIATION OF MARIJUANA USE AMONG CHILDREN AND PARENTS

5.1 Prevalence of Marijuana Use Among Children and Parents

Prevalence rates of marijuana use among parents and children in the total dyads and by sociodemographic characteristics are presented as background to the analysis of association in patterns of use within dyads. Data for 12-17 year olds are available in all surveys but only in 1991-1996 for 18-25 year olds. To adjust for the variable age distributions of children in dyads from different surveys, age standardized estimates were calculated separately among 12-17 year olds for 1979-1996 and among 18-25 year olds for 1991-1996 based on the 1991 distribution of child age. Data for the aggregated 1979-1996 surveys are presented in Tables 5.1 and 5.2 and for specific years in Appendix Tables A.5.1 and A.5.2. Use among children and their parents follow well described age, gender and ethnic related patterns.

5.1.a Child Patterns of Marijuana Use

In the 1979-1996 aggregate sample, 15.1% of children aged 12-17 had ever used marijuana and 11.9% had used it in the last year (Table 5.1). In 1991-1996, 10.8% and 8.4% of 12-17 year olds had done so; 40.1% of youths aged 18-25 had used marijuana in their lifetime and 23.7% in the last year.

The increasing rates of use with increasing age followed well documented age related trends. Adolescents were differentiated into two age groups, 12-14 and 15-17 years old. In 1991-1996, there was a five-fold increase in rates of lifetime marijuana use among older adolescents aged 15-17 years (19.2%) compared with younger adolescents aged 12-14 years (3.2%), and a further doubling from ages 15-17 to ages 18-25 (40.1%).

While in the total sample of 12-25 year olds, boys were more likely than girls to use marijuana lifetime and in the last year (p's<.001), this sex difference did not appear among adolescents aged 12-14 and 15-17 for lifetime and last year use (see Figure 5.1). The interaction term of child age and sex was statistically significant only for last year use (logistic regression, p<.05).

There were no statistically significant ethnic differences in rates of lifetime and last year marijuana use in the total sample of youths.

Figure 5.1. Prevalence of Lifetime and Last Year Marijuana Use Among Children Aged 12-25, by Child Age and Sex^1,2 (NHSDA 1979-1996)

¹Weighted estimates.
²SUDAAN PROC LOGISTIC (Wald F-test) was used to test for age and sex differences in lifetime and last year marijuana use.
^a-dFor child last year marijuana use, percentages with different superscripts are significantly different from each other.

5.1.b Parental Patterns of Marijuana Use

In the aggregate 1979-1996 sample, 32.8% of parents had ever used marijuana in their lifetime but only 5.3% had used in the last year (Table 5.2). Despite the large number of dyads in the sample, this low prevalence rate resulted in a relatively small number of families (N= 572) in which the parents reported to be still using marijuana within the last year (343 mothers, 229 fathers). Parents who reported using marijuana within the last year preceding the interview reported greater lifetime usage than those who had stopped using marijuana. In the 1979-1994A aggregated surveys, 41.2% of current users reported having ever used marijuana 100+ times compared with 17.7% of former users (Appendix Table A.5.11). Lifetime parental rates increased from 1979 to 1994, peaked at 47.9% in 1994A (36.3% in 1994B), and declined in 1995 and 1996. Parent last year marijuana use remained stable during the 17-year period covered by the surveys (Appendix Table A.5.2). The increase in lifetime prevalence over most of the period covered by the surveys reflects historical changes in rates of marijuana use initiation experienced in their adolescence by parents from different birth cohorts.

Fathers had higher rates of marijuana use than mothers: 38.9% of fathers and 28.7% of mothers reported having used marijuana in their lifetime; 7.1% of fathers and 4.1% of mothers reported using it in the last year (p's<.001) (Table 5.2). This pattern was very stable across individual survey years (Appendix Table A.5.2).

Parental lifetime marijuana use decreased with increasing child age as parental age increased.

There were significant ethnic differences in patterns of parental marijuana use. African-American and white parents had higher rates of lifetime marijuana use than Hispanic parents (p's<.001); African-American parents had higher rates of last year use than white and Hispanic parents (p's<.001) (Table 5.2). The different ethnic patterns observed among children and their parents mirror the epidemiological finding that, in adolescence and early adulthood, whites usually have higher rates of drug use than minorities, while the ethnic patterns reverse beginning in middle adulthood.

5.1.c Drug Use Among Respondents in Dyads and Those Not in Dyads

We noted earlier that the dyads in the NHSDA were not selected to constitute a representative sample of parent-child pairs in the United States, and that the sociodemographic characteristics of parents and children in dyads differed somewhat from those not in dyads. Two features of the epidemiology of marijuana use differed in the two groups as well, especially for dyads with the youngest adolescents 12-14 years old.

We compared the rates of self-reported marijuana use by children and parents in dyads with youths of similar ages and parents not included in dyads. The rates of lifetime and past year marijuana use among 12-14 year olds and of lifetime use among 18-25 year olds were generally significantly higher in the non-dyadic sample than in the dyadic sample of children (Table 5.3). Rates of use for parents not in dyads, controlling for child age, were generally the same in dyads and non-dyads. The only statistically significant difference was the higher lifetime rates of parents of 12-14 year olds not in dyads (Table 5.3). Detailed year-specific data are presented in Appendix Tables A.5.3 and A.5.4.

Rates of use over time in the sample of youths in dyads fluctuated more and rates for 18-25 year olds deviated somewhat from historical trends observed among youths in the United States during the same historical period. This can be seen from a comparison of trends in last year prevalence of use among youths aged 12-17 and 18-25 for the total NHSDA (Figure 1.1) and for the dyads (Appendix Figure A.5.1). Rates of marijuana use among 18 to 25 year olds in the dyads started to decline as of 1995 at a time when, in the U.S. as a whole, rates of use in that age group were still increasing.

5.2 Association in Marijuana Use Between Parents and Children

The influence of parents on children was inferred from the extent of association in marijuana use between them. As noted earlier, although the data are cross-sectional, we use the term parental influence, since we assume that it is unlikely that parents use marijuana in response to their children's use.

The association between children's and parents' patterns of marijuana use within dyads was estimated by odds ratios between child and parent use. Cross-tabulations for the aggregated 1979-1996 surveys are presented in Table 5.4 and odds ratios in Table 5.5. Cross-tabulations for each of the 10 surveys separately are presented in Appendix Tables A.5.5-A.5.9 (A.5.5 by child age, A.5.6 by child sex, A.5.7 by parent sex, A.5.8 by dyad type, A.5.9 by ethnicity) and odds ratios in Appendix Table A.5.10.

5.2.a Cross-Tabulations Between Parent and Child Marijuana Use

Parental lifetime and last year use. In a first step, the children's lifetime and last year marijuana use was examined as a function of parental history of marijuana use. There were consistently positive and significant associations between the marijuana behavior of parents and children (Table 5.4). In the total sample of youths aged 12-25, the proportion reporting to have ever used marijuana was more than forty percent higher when parents reported to have ever used marijuana in their lives (22.9%) compared with when parents had never used (15.6%). Rates of last year use by the children were sixty percent higher (17.5% and 11.0%, respectively, in each type of family). When parents reported using marijuana within the last year almost twice as many (21.9%) of the children had also used marijuana in the same period compared with when parents had not used marijuana (12.7%).

Parental lifetime and last year use were considered simultaneously in order to distinguish the influence of former from current use (i.e., use within the last year). In the total sample, the percentages of children using marijuana, whether lifetime or last year, were similar for former and current parental marijuana users and not statistically significantly different from each other (Table 5.4). When the sample was disaggregated by age, however, a difference appeared among the youngest adolescents aged 12-14. There were higher percentages of lifetime and past year users among the children of current than of former marijuana users (differences statistically significant at P<. 05).

5.2.b Measures of Association Between Parent and Child Marijuana Use

The odds ratio, a measure of association that is insensitive to marginal distributions, provides a more appropriate assessment of parent-child association than the percentages of dyad members with similar patterns of use. In a first step, univariate regression models were estimated to predict the child lifetime and last year marijuana use as a function of parental pattern of marijuana use, without control for any other variable. Models were run with different definitions of parental marijuana use: lifetime, last year, combined lifetime-last year (Table 5.5).

The influence of lifetime and last year parental use was similar. The unadjusted odds of children using marijuana ever or within the last year when their parents were lifetime or last year users were almost identical (1.6-1.9). The odds varied slightly from year to year (Appendix Table A.5.10). The highest odds appeared in the very small 1994A sample and reflected imprecision of the estimates. The joint examination of lifetime and last year use confirmed the descriptive finding reported above regarding the relative impact of current versus former parental use on lifetime use by the child. In the total sample, the effects were not significantly different (Table 5.5). Although, compared with never users, last year parental users were more likely (OR=2.3, p<.001) than former users (OR=1.6, p<.001) to have a child who reported using marijuana in the last year, the difference between the odds was not statistically significant. As discussed above with respect to rates of use, the patterns changed somewhat by children's age and ethnicity. Among the youngest children and Hispanics, parental current use had a stronger effect than former use (differences in the odds ratios significant at p<.01).

Parental influence was apparent for children of all ages (Table 5.5). The odds in the age-specific groups were consistently higher than in the sample as a whole, reflecting the association of child age with levels of marijuana use. The association of parental lifetime use with child lifetime use was significantly higher for the oldest children than the two younger groups (p<.05; Wald test); the association of parental last year use with child lifetime use was significantly higher both for the oldest and youngest offspring than those 15-17 years old (p's<.05; Wald test). Parental influence displays a curvilinear pattern, declining throughout adolescence and increasing among young adults aged 18 to 25. This probably reflects the closeness and dependence on parents of these older youths who are still living at home, although some are married.

The association of children's marijuana use with mothers was stronger than with fathers, especially when the parent was still using marijuana within the last year preceding the interview (Table 5.5). However, the differences were not statistically significant.

Parental lifetime marijuana use had statistically significant and similar associations with boys' and girls' lifetime and last year use (Table 5.5). Parental use in the last year, however, was significantly associated with the lifetime and last year use of sons only (ORs=2.3,2.2 p's<.001), but the differences between boys and girls were not significant.

Trends appeared in same- and cross-sex patterns of association between parents and children, but they were not statistically significant (Table 5.5). When fathers smoked, the odds of smoking by children were elevated only among sons; when mothers smoked, the odds were elevated among sons and daughters. For example, when fathers used marijuana in the last year, sons were more than twice as likely to use marijuana both in their lifetime and in the last year (ORs=2.1 and 2.0, p's <.05) while daughters were not (OR=1.1). When mothers used marijuana in the last year, the odds ratios for sons' and daughters' last year use were 2.5 (p<.01) and 2.0 (p<.05), respectively. However, the three-way interactions between parent sex, child sex and parent marijuana use on child marijuana were not statistically significant (data not presented). Consequently, all multivariate analyses were implemented without differentiating parents and children by gender.

Across ethnic groups, parental lifetime marijuana use was positively and significantly associated with child lifetime and last year marijuana use (Table 5.5). Ethnic differences in patterns of association appeared only with respect to parental last year use. The association of parental last year marijuana use with child lifetime and last year marijuana use was higher among Hispanics (ORs=3.3, 4.6, p's <.001) than whites (ORs=1.7, 1.8, p's <.05) and African-Americans (OR = 1.5, ns; 1.7, p <.05). The unadjusted odds ratios of parental last year marijuana use on child marijuana use among Hispanics were slightly more than twice those of whites and African-Americans for lifetime use and almost three times those for last year use (p's<.05; Wald test). However, the ethnic differences become non-significant with control for sociodemographic covariates (see Chapter 6).

5.2.c Extensiveness of Parental Marijuana Use and Child Marijuana Use

Three measures of parental extensiveness of marijuana use were available: total number of days (or times) used lifetime, past year and past month. Because of the relatively small number of parents who were using within the past month, the analysis of children's marijuana use as a function of the extensiveness of very recent parental use could only be exploratory.

The format of the frequency of lifetime use varied across survey years. In particular, the highest number of days used was expanded from a maximum of 100+ to 200+ and then to 300+ (see Technical Appendix). Since children's use did not vary as a function of these three categories (Table A.5.12), a four-category variable was used across the 10 surveys, where 100+ was the highest category. Percentages of children's using marijuana as a function of extensiveness of parental use and odds ratios are displayed in Table 5.6. There is a slight trend for the percentages of marijuana using children to increase as a function of extensiveness of parental use. However, none of the differences were statistically significant. There were no significant variations in children's lifetime or last year use by extensiveness of parental use. The high odds observed for the missing category among parents who used marijuana in the last year remains to be explained, especially since these missing cases were structurally missing and did not reflect any self-selection bias on the part of participants in the survey.

While we expected parental influence to vary as a function of recency and extensiveness of use, no such effects were found. Overall parental influence is moderate. As will be seen in the next chapter, the odds between parental and child marijuana use increase when other factors are controlled, in particular child and parent age.

Table 5.1. Prevalence of Child Lifetime and Last Year Marijuana Use Among Children Aged 12-25^1,2 in Parent-Child Dyads, by Child Age, Sex and Ethnicity (NHSDA 1979-1996)

		1979- 1996		1979-1990		1991-1996
		%		%		%
Lifetime Use
Child Age3
12-14 years		5.4	a	11.9	a	3.2	a
15-17 years		26.7	b	44.1	b	19.2	b
18-25 years4		40.1		-		40.1
12-17 years		15.1		27.3		10.8
Child Sex
Male		19.4	a	27.8		16.8	a
Female		16.5	b	27.2		13.6	b
Child Ethnicity
White		18.8		28.7		15.3
African-American		16.4		25.3		14.6
Hispanic		17.0		19.8		16.5
Last Year Use
Child Age3
12-14 years		4.4	a	9.7	a	2.7	a
15-17 years		20.8	b	35.1	b	14.7	b
18-25 years4		23.7		-		23.7
12-17 years		11.9		21.8		8.4
Child Sex
Male		14.9	a	22.2		12.6	a
Female		11.4	b	21.8		8.5	b
Child Ethnicity
White		13.9		22.9		10.7
African-American		12.1		21.0		10.3
Hispanic		11.9		15.9		11.2
Total N		9,463		1,538		7,925

¹ Weighted estimates with SUDAAN PROC CROSSTAB, unweighted N's.
² For 1979, 1982, 1988 and 1990, children aged 12-17 were selected. For all other years, children aged 12-25 were selected.
³ Adjusted estimates based on the 1991 distribution of child age for 12-17 and 18-25 year olds.
⁴NHSDA 1991-1996.
^a-bFor each sample and each sociodemographic variable, percentages with different superscripts are significantly different from each other, T-test (p<.05).
Source: SAMHSA, Office of Applied Studies, National Household Survey on Drug Abuse.

Table 5.2. Prevalence of Parent Lifetime and Last Year Marijuana Use in Parent-Child Dyads, by Child/Parent Age, Sex and Ethnicity^1,2 (NHSDA 1979-1996)

	1979-1996		1979-1990		1991-1996
	%		%		%
Lifetime Use
Total Parents	32.8		17.8		37.2
Child/Parent Age
12-14/24-73 years	37.5	a	23.2	a	42.2	a
15-17/27-80 years	29.8	b	12.2	b	36.2	b
18-25/31-74 years	24.2	c	-		24.2	c
Parent Sex
Male	38.9	a	20.9		44.7	a
Female	28.7	b	15.4		32.3	b
Parent Ethnicity
White	34.7	a	16.3	a	41.2	a
African-American	37.3	a	31.6	b	38.5	a
Hispanic	20.5	b	8.3	c	22.7	b
Last Year Use
Total Parents	5.3		6.9		4.8
Child/Parent Age
12-14/24-73 years	7.0	a	9.7	a	6.1	a
15-17/27-80 years	3.9	b	3.8	b	3.9	b
18-25/31-74 years	3.2	b	-		3.2	b
Parent Sex
Male	7.1	a	8.7		6.6	a
Female	4.1	b	5.5		3.7	b
Parent Ethnicity
White	4.8	a	5.6	a	4.5	a
African-American	8.5	b	15.1	b	7.1	b
Hispanic	3.5	a	4.1	a	3.5	a
Total N	9,463		1,538		7,925

¹ Weighted estimates with SUDAAN PROC CROSSTAB, unweighted N's.
² For 1979, 1982, 1988 and 1990, children aged 12-17 were selected. For all other years, children aged 12-25 were selected.
^a-cFor each sample and each sociodemographic variable, percentages with different superscripts are significantly different from each other, T-test (p<.05).
Source: SAMHSA, Office of Applied Studies, National Household Survey on Drug Abuse.

Table 5.3. Prevalence of Parent and Child Lifetime and Last Year Marijuana Use Among Parents and Children Aged 12-25 by Membership in Parent-Child Dyads^1,2 (NHSDA 1979-1996)

	Children				Parents
			Non-				Non-
	Dyads3		Dyads		Dyads3		Dyads
	%		%		%		%
Lifetime Use
Child/Parent Age4,5
12-14/24-87years	5.4		6.8	***	37.5		44.6	***
15-17/26-80 years	26.7		26.8		29.8		38.2
18-25/26-74 years	40.1		45.4	***	24.2		24.8
12-17/24-87 years	15.1		17.1	***	36.5		39.9
Last Year Use
Child/Parent Age4,5
12-14/24-87years	4.4		5.5	**	7.0		6.8
15-17/26-80 years	20.8		21.3		3.9		4.4
18-25/26-74 years	23.7		23.0		3.2		2.6
12-17/24-87 years	11.9		13.6		5.6		5.8
Age Specific N's4,5
12-14/24-87years	4,794		18,295		4,223		3,078
15-17/26-80 years	3,598		19,519		3,105		2,779
18-25/26-74 years	1,070		33,056		1,070		3,645
12-17/24-87 years	8,392		37,814		8,392		6,083
Total N5	9,462		70,870		9,462		9,728

¹ Weighted estimates, unweighted N's.
² In 1979, 1982, 1988 and 1990, children aged 12-17 were selected. In all other years, children aged 12-25 were selected.
³ For parent-child dyads, adjusted estimates based on the 1991 distribution of child age for 12-17 and 18-25 year olds.
⁴ Age ranges for dyad parents are 24-73, 27-80 and 31-74; and for non-dyad parents 26-87, 26-71 and 26-71, for children aged 12-14, 15-17 and 18-25, respectively.
⁵ Differentiation of children aged 12-14 and 15-17 not available for non-dyad parents in 1979 and 1982.
Estimates of dyad and non-dyad parental marijuana use for 12-14 and 15-17 year olds excludes the 1979 and 1982 surveys; estimates for 12-17 year olds includes all survey years. Data for 18-25 year olds available in NHSDA 1991-1996.
* p<.05; ** p<.01; ***p<.001, Z-test of the percentage difference between dyad and non-dyad parents and children.
Source:SAMHSA, Office of Applied Studies, National Household Survey on Drug Abuse.

Table 5.4. Lifetime and Last Year Marijuana Use Among Children Aged 12-25^1,2 by Parent Pattern of Use and Child Age (NHSDA 1979-1996)

			Parent Marijuana Use
			Lifetime				Last Year				Current
Child			No		Yes		No		Yes		Never		Former		Last Year
Marijuana Use			%		%		%		%		%		%		%
Total
Lifetime			15.6		22.9	***	17.5		26.9	**	15.6	a	22.2	b	26.9	***b
Last Year			11.0		17.5	***	12.7		21.9	**	11.0	a	16.7	b	21.9	***b
N			6,379		3,084		8,891		572		6,379		2,512		572
Children Aged
12-14 years
Lifetime			3.5		8.3	***	4.5		15.6	**	3.5	a	6.6	b	15.6	***c
Last Year			2.9		6.8	***	3.7		12.8	**	2.9	a	5.4	b	12.8	***c
N			3,020		1,774		4,443		351		3,020		1,423		351
15-17 years
Lifetime			22.5		35.1	***	25.8		38.0	*	22.5	a	34.7	b	38.0	***b
Last Year			17.1		28.2	***	19.9		33.2	*	17.1	a	27.4	b	33.2	***b
N			2,529		1,069		3,410		188		2,529		881		188
18-25 years3
Lifetime			33.5		63.0	***	39.3		80.4	*	33.5	a	60.3	b	80.4	***b
Last Year			18.5		38.9	***	22.4		54.5		18.5	a	36.5	b	54.5	**b
N			830		241		1,038		33		830		208		33

¹ In 1979, 1982 and 1990, children aged 12-17 were selected. In all other years, children aged 12-25 were selected.
² Weighted estimates with SUDAAN PROC CROSSTAB, unweighted N's.
³ NHSDA 1991-1996.
^a-c Comparisons across categories of use for each pattern of use: percentages with different superscripts are significantly different from each other, Wald F-test (p<.05). *p<.05; **p<.01; ***p<.001, X² test.
Source: SAMHSA, Office of Applied Studies, National Household Survey on Drug Abuse.

Table 5.5. Association in Marijuana Use Between Parents and Children Aged 12-25¹, by Child Age, Sex, Ethnicity, Parent Sex and Parent-Child Dyad Type, Unadjusted Odds Ratios² (NHSDA 1979-1996)

		Parent Lifetime Marijuana Use
		Child				Child
Child Marijuana Use	N	Lifetime		(95% CI)		Last Year		(95% CI)
Total	9,463	1.6	***	(1.4-1.9)		1.7	***	(1.4-2.1)
Children Aged
12-14 years	4,794	2.5	***	(1.7-3.7)	ab	2.4	***	(1.5-3.8)
15-17 years	3,598	1.9	***	(1.5-2.4)	a	1.9	***	(1.5-2.5)
18-25 years3	1,071	3.4	***	(2.2-5.2)	b	2.8	***	(1.7-4.6)
Child Sex
Son	4,807	1.6	***	(1.3-2.0)		1.7	***	(1.3-2.2)
Daughter	4,656	1.7	***	(1.3-2.1)		1.8	***	(1.3-2.3)
Parent Sex
Father	2,922	1.5	**	(1.1-2.1)		1.5	*	(1.0-2.0)
Mother	6,541	1.7	***	(1.4-2.1)		2.0	***	(1.6-2.4)
Parent-Child Dyad Type		</