risk and protective factors for alcohol and other drug problems in adolescence and early adulthood:...

7/29/2019 Risk and Protective Factors for Alcohol and Other Drug Problems in Adolescence and Early Adulthood: Implications

1/43

!!!!!

!!!!!!!!!!!!!!!!!!!!!P.!O.!Box!6448!!Reno,!NV!89513!!Phone:[email protected]!

Title&Below&please&list&the&title&of&this&resource.&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&

!

Risk!and!Protective!Factors!for!Alcohol!and!Other!Drug!Problems!in!Adolescence!and!Early!

Adulthood:!Implications!for!Substance!Abuse!Prevention!!!

!

Author&Below&please&list&the&author(s)&of&this&resource."

!

J.!David!Hawkins,!Richard!E!Catalano,!and!Janet!Y!Miller!!!Citation&Below&please&cite&this&resource&in&APA&style.&For&guidance&on&citation&format,&please&visit&

http://owl.english.purdue.edu/owl/resource/560/01/&

!

Hawkins,!J.,!Catalano,!R.,!&!Miller,!J.!(1992).!Risk!and!Protective!Factors!for!Alcohol!and!

Other!Drug!Problems!in!Adolescence!and!Early!Adulthood:!Implications!for!Substance!Abuse!Prevention.!Psychology"Bulletin,!112,1,!64105.!

!!

Summary&Below&please&provide&a&brief&summary&of&this&resource.&If&an&abstract&is&available,&feel&free&to&copy&and&paste&it&here.&

!

The!authors!suggest!that!the!most!promising!route!to!effective!strategies!for!the!prevention!of!adolescent!alcohol!and!other!drug!problems!is!through!a!riskfocused!approach.!This!

approach!requires!the!identification!of!risk!factors!for!drug!abuse,!identification!of!

methods!by!which!risk!factors!have!been!effectively!addressed,!and!application!of!these!

methods!to!appropriate!highrisk!and!general!population!samples!in!controlled!studies.!

The!authors!review!risk!and!protective!factors!for!drug!abuse,!assess!a!number!of!

approaches!for!drug!abuse!prevention!potential!with!highrisk!groups,!and!make!

recommendations!for!research!and!practice.!!

!

!

!

!!

!

!

!"


2/43

Psychological Bul le t in1992, Vol. 112 . No. 1.64-105 C o p \ r i g h t 1992 h> t h e Amer ican Psychologica l Assoc ia t ion , I I

Risk and Protective Factors for Alcohol and Other Drug Problemsin Adolescence and Early Adulthood: Implicationsfor Substance Abuse PreventionJ. David Hawkins, Richard E Catalano, and Janet Y MillerSocial Development Research Group, School of Social WorkUniversity of Washington

The authors suggest that the most promising route to effective strategies for the prevention ofadolescent alcohol and other drug problems is through a risk-focused approach. This approachrequires the identification of risk factors for drug abuse, identification of methods by which riskfactors have been effectively addressed, and application of these methods to appropriate high-riskand general population samples in controlled studies. The au thors review risk and protective factorsfor dru g abuse, assess a number of approaches for drug abuse prevention potential with high-riskgroups, an d make recommendations for research an d practice.

In spite of general decreases in the prevalence of the nonmed-ical use of most legal and illegal drugs in recent years, the abuseof alcohol and other drugs during adolescence and earlyadulthood remains a serious public health problem (Adams,Blanken, Ferguson, & Kopstein, 1990). The consequences ofdrug abuse are acute on both a personal and a societal level.Forthe developing young adult, drug and alcohol abuse under-mines motivation, in terferes with cognitive processes, contrib-utes to debilitating mood disorders, and increases risk of acci-dental injury or death. For the society at large, adolescent sub-stance abuse extracts a high cost in health care, educationalfailure, me ntal health services, drug an d alcohol treatment, an djuvenile crime.Added to the immediate personal and social costs of adoles-cent drug abuse are the longer range implications for young-sters who continue to abuse alcohol and drugs into adult life.Drug abuse is involved in one third to one half of lung cancerand coronary heart disease cases in adults (R. Blum, 1987).Alcohol and other drugs are major factors in acquired imm uno -deficiency syndrome (AIDS), violent crimes, child abuse andneglect, and un em ploy men t. The problems associated with al-cohol and othe r drug abuse carry costs in lost p roductivity, lost

Prepar ation of this article was supported in part by Grant DA03721from the National Institute on Drug Abuse and by Grant 87-JS-CX-K084 from the Office of Juvenile Justice and Delinquency P revention.Points of view or opinions expressed are those of the authors.Our thanks to George S. Bridges, Elise Lake, Randy Gainey, TimMurphy, John Campos, and Cheryl Yates for their assistance in thepreparation of this article and to Patricia Huling for her managementof the document's creation. Thanks also to Barry Brown, DavidFarrington, Michael Goodstadt, Rolf Loeber, Roger Weissberg, He-lene White, and two anonymous reviewers for their comments ondrafts.Correspondence co ncernin g this article should be addressed to J.David Haw kins, Social Development Research Group , School of So-cial Work, University of Washington, 146 N. Canal Street, Suite 211,Seattle, Washington 98103.

life, destruction of families, and a weakeningof the bonds thathold the society together.Given the serious consequences of drug and alcohol abuse,considerable effort has been directed toward identifying effec-tive treatment. Until recently, applied research in the substanceabuse field has consisted primarily of experimental trials ofvarious forms of treatment for alcohol an d other drug abuse.The goal has been to identify ways to increase the effectivenessof treatmen t and to preven t relapse following treatment. Strate-gies ranging from self-help to aversive counterconditioninghave been advocated and assessed.Many of these studies have demonstrated how abstinence

can be achieved, but long-term mainten ance of abstinence hasbeen more difficult. Th e reinforcing properties of alcohol andother drugs are themselves often reinforced by norms and be-haviors of family members and others in the communities inwhich recovering people live. These combined reinforcementsoften overcome short-term treatment gains. According to thesurgeon general: "For m any drug-depend ent persons, achiev ingat least brief periods of dru g abstinence is a readily achievablegoal. Maintaining abstinence, or avoiding relapse, however,poses a much greater overall challenge" (Surgeon general, 1988,P. 311) .Added to disappointment with th e staying power of drugtreatmen t is a growing recognition of the high cost of treatmentan d of the inability of existing treatment programs to keep upwith increasing demand . In recent years, these considerationshave stimulated interest in primary prevention of alcohol andother drug abuse.This article focuses on the prevention of alcohol and otherdrug abuse among adolescents. A number of views have beenadvanced about what constitutes substance abuse when con-sidering adolescents (Hawkins, Lishner, & Catalano, 1985). Inthis article, adolescent drug abuse isdenned as the frequent useof alcohol or other drugs durin g the teenage years or the use ofalcohol or other drugs in a manner that is associated with prob-lems an d dysfunctions. This conception of the problem is not

64


3/43

RISK AN D PROTECTIVE FACTORS FO R DRUG PROBLEMS 65meant to condone the infrequent use of alcohol or other drugsby teenagers, which is a violation of the law. The present defini-tion simply reflects a recognition that a relatively large propor-tion of teenagers try alcohol or other drugs without becominginvolved in the frequent use of these substances or developingdrug-related problems (Newcomb & Bentler, 1988; Shedler &Block, 1990).

Cloninger and his colleagues (Cloninger, Bohman, Sigvards-son, & von Knorring, 1985; Cloninger, Sigvardsson, & Boh-man, 1988) have identified two typesof alcoholism. One type isassociated w ith frequ ent impulsive-aggressive behavior and fol-lows an early onset of alcohol use and alcohol problems in ado-lescence. This type of drug abuse is considered in this article. Itis distinct from alcoholism that develops after age 25, which isnot a focus of the current article.Precursors of drug and alcohol problems have been de-scribed as risk factors for drug abuse. Risk factors occur beforedrug abuse and are associated statistically with an increasedprobability of drug abuse. A risk-focused approach seeks toprevent drug abuse by eliminating, reducing, or mitigating itsprecursors. This article suggests that a promising line for pre-vention research lies in testing interventions targeting mu ltipleearly risk factors for drug abuse.A risk-focused approach in drug abuse prevention researchand policy is warranted given the apparent success of this ap-proach in reducing risk factors for problems as divergent asheart and lung disease (Bush et al., 1989; Vartiainen, Pallonen ,McAlister, & Puska, 1990) and school failure (Berrueta-Cle-ment, Schweinhart, Barnett, Epstein, & Weikhart, 1984). Theapparent failure of early prevention interve ntions, such as druginformation programs that did not address known risk factorsfor drug abuse (Stuart, 1974; Weaver & Tennant, 1973), alsoargues for this approach.Many of the risk factors for adolescent drug abuse also pre-dict other adolescent problem behaviors (Hawkins, Jenson, Ca-talano, & Lish ner, 1988). There is evidence that adolescent drugabuse is correlated with delinquency teenage pregnancy, andschool misbehaviorand drop out (Ellio tt, Huizinga, & Menard,1989; Jessor & Jessor, 1977; Zabin, Hardy, Smith, & Hirsch,1986). Comprehensive risk-focused efforts probably can pre-vent other adolescent problem behaviors besides drug abuse.If prevention of drug abuse (asdenned above) is the goal, th enrisk factors salient for drug abuse rather than for the occasionaluse of alcohol or other dru gs should be targeted. A relativelysmall propo rtion of adolescent d rinke rs or users are frequent orproblem users (Johnston, O'Malley, & Bachman, 1988; Shedler& Block, 1990). The following review focuses on factors thathave been show n to precede dru g abuse.

Risk Factors for Adolescent Drug AbuseMost studies to date have focused on small subsets of identi-fiable risk factors for drug abuse. There is little evidence avail-able regarding the relative importance and interactions ofvarious risk factors in the etiology of drug abuse, althoughcurrent studies are seeking to measure a broader range of iden-tified risk factors. At this time, it is difficult to ascertain, forinstance, which risk factors or combination of risk factors aremost virulent, whic h are m odifiable, and which are specific to

drug abuse rather than generic contributors to adolescent prob-lem behaviors. Current knowledge about the risk factors fordrug abuse does not provide a formula for prevention, but itdoes point to potential targets for preventive intervention. Im-plications for intervention are considered in this article after areview of know n risk factors for drug abu se in adolescence andearly adulthood.These risk factors can be roughly divided into two categories.First are broad societal and cultural (i.e., contextual) factors,which provide the legal and normative expectations for behav-ior. The second group includes factors that lie wi thin individ-uals and their interpersonal en vironments. The principal inter-personal environ ments in children's lives are fa milies, schoolclassrooms, and peer groups. The risk factors have been de-scribed elsewhere (Hawkins, Lishner, Catalano, & Howard,1986; Kandel, Simcha-Fagan, & Davies, 1986; Newcomb, Mad-dahian, & Bentler, 1986; Simcha-Fagan, Gersten, & Langner,1986) and are summarized here and in the left half of Table 1.This is not intended as a critical review of the methodologies ofthe studie s but rather as an overview of the evidence curren tlyavailable on risk factors for adolescent d rug abuse.

Contextual FactorsIndividuals and g roups exist within a social context: the val-ues and structure of the ir society. For example, shifts in culturalnorms, in the legal definitions of certain behaviors, and in eco-nomic factors have been shown to be associated with changes indrug-using behaviors and in the prevalence of drug abuse. Thefollowing risk factors (1 through 4 below) exist in the broadsocial context:1. Laws and norms favorable toward behavior. Recent re-search on the effects of laws on alcohol consumption has fo-cused on three interventions by law: (a) taxation, (b) laws stating

to whom alcohol may be sold, and (c) laws regarding how alco-hol is to be sold.Alcohol consumption is affected by price, specifically theamount of tax placed on alcohol at purchase (Lev y & Sheflin,1985). Cook and Tauchen (1982) found that increases in taxeson alcohol led to immediate and sharp decreases in liquor con-sumption and cirrhosis mortality.Studies examining the relationship of minimum drinkingage and adolescent dri nk ing and dr ivin g have generally shownthat lowering the drinking ag e increases teen drinking an ddriving and teen traffic fatalities and raising it decreases teendriving while intoxicated citations (DWIs) and deaths (Cook &Tauchen, 1984; Joksch, 1988; Krieg, 1982; Saffer & Grossman,1987).Studies of restriction on how alcohol is sold have shown thatallowing patrons to purchase distilled spirits by the drink in-creased the consumption of distilled spirits and the frequencyof alcohol-related car accidents (Holder & Blose, 1987). How-ever, there was no increase in accidents involving males u nderthe legal drink ing age of 21 (Blose & Holder, 1987).Two general explanations of how laws affect the use of sub-stances have been advanced. The first posits that laws reflectsocial norms and that use is largely a function of group norms(Watts & Rabow, 1983). Alcohol consumption rates vary among(text continues on page 81)


4/43

66 D. HAWKINS, R. CATALANO, AND J. MILLER

&c^

pnP

oFin,3 i .T>-p 5 u a c cfliil Hilllllll 111 11B 3liL- ^j 2 X! to t^8 j f i g S 5 S ^2.g[3j< e B o 20 S ^8S ^^j 5>"* i> -$* ~tt"j~ 0

|1 io

r"33 u< ^ ^Sj "^* 0-.E .-Sc->

-a ooc .grt 'KS o -s|! 'S -a 3 -3 *tH IS i.s 8 s g Q o rf S3a tf-e a-so a g o u21-3 - s aS u U 4 2 2 ** S? r- t>

^ S g = igt sS^s oo 3 to o g.,3.S -o -a > B 3 S 3 S S ,23 J? 2 S H Po ? la o >rs-Ml'ls ! 2 - 8"B C T> 0 _0 5)Z w.- 852 waa 2| U-

S 2? ;5 ^ Os 1"

s il a 131 -S^ aB a3- 3c 3 1, 3o s as ~ S

Bos1

0000o\

saaT3

aoaewh

T-00OSQ1

_00

1~< IO3 ur^ "r^C T3 00 vf . ". -~' i l i lp^llor t 2 ^ en- o ^ cC ^ ^ r t - C i o W5 .1 'S oo a--e -S 7 1 S1 c t a g g 1 " 3 ^ g - 5 |.>'0^ S g|'SJ? U

O(3Qs s^"? u oo2 oa *

JK" if rJ g CQ CU

00M .S oo

soomacfao

ehcaohgonm

[oeeueoapomo

acameren

hgdnnes

2

Tf* ~ 00 JJ 0

~tn ^ W.sis"SI eo i > 5e So BPS

*sss2o~lo BII


5/43

RISK AN D PROTECTIVE FACTORS FOR DRUG PROBLEMS 67stc

1

t O\\O \OOO OO OOa\ o\ o\ - 2I

nocdua

w

ojsi

1"o"oooq,1

IS

(-o^J(2y;1

_ *-*^J ug a9 O 4> r^ ** cto JS on oc Cg ' g ^ S ^^ rt ^

SUS as> Si 8."73 W "138 g H, a &

r-^

aesksannmpo

famiyneaore

cdpoembo

cu

oooo _"> .< 00"-73 g 2: - 5 -? "(g'coo e m ST^ S fs" oo h oo i-*0 g < * 0- S"-11"(X

o S!i*

vnvmebo

ampebdsa

pmsvbh

aoaewhrsk

O

=8* ^\__r ed t~-S ^

1 J 9 -!fc^ &2 3N

i| f - s _ e g Hi^I l !iln illl 1f i l l 111!hf i to o B 9 .j- g C ioS-?-o5 l - g . 2 3 o o ^ S ^ o . 2 , T = S , 8 t ; S0 8 e g -S e > ! S 8 .S o 52, s St 8 3S ^ ^ - S S K . S 3 . 3 S S ' S J ^ c S ' o ' eX S3 5 29 -a ^,cK _ o s =g -a11 1 -sl s ll 5 g t g i - t i W (5 UK t-i Jj 4J oj CO 'C P ^ C UiLl !!sf!5 lit Hi!SSilIiilllilil5||.^i ll|illilliililllhli 11 a i( S o Q o-

>Crn ooo .O Oi 1 Is ?1 * -.

22 -i ^ 8S oi z H

hmewkagmes

Pencmp

n

a

saeyb

poamp

aoae

whoweraeoo

ac

amauu


9/43

RISK AN D PROTECTIVE FACTORS FOR DRUG PROBLEMS 71

sia

8Ills-d f a

8CTsa.X

. -g

1-OO

>>1 BII


10/43


G 3O f,

'

E

jj' oo "O r

^ e _ C Mt s j . 2 ' 8 8 S^^ 3'i3 l'S-3Slllllllll

"^ 3-

w-4~>Cfll

empami

Dc"

oo3V*T30C

cbe

yod

T3O

inah

poemi

CTJGca3'CagcCD3CT>

00

C"Wiuo

emoo

1z1 cd

r i^f)4J>i/S>15_CC/51/5Aevm

U&n3ift3"HOX5t->t"3/~\

i- S.0 S> a JS crjCU U2 >>(U ^

.> 015Q"">T3poema

manan

CQ^ 9 _*?


11/43

RISK AND PROTECTIVE FACTORS FOR DRUG PROBLEMS 73

Sisa

B

S? "O i_ t> j- u J3ilils.1

!1!fa u * .S 2c ^U ^T3 toS'-au

tr o' E T> s "5 n. v '"^11.1111X K

J2 BS 3In

sj-o^jS " l o g . ,h U o f>


12/43


f3g;SOSU

1

a.0cDVB

0 ^_,S | c

Efesosa

ua

n

XT355

encots_o"S.eH

3s5

;enn

ug."2'>w

xatn173iS

ots"'-- S -P g o |-g^ 1 2 s "S S |1 g s a ^*I1-|*= si; s^ > ^ > o s1111 1 -!s||^f lst l ilaillw o^ M ' s ; J _ * j 03 . S < 5 * -^* "8l^K^| g ai ^ r u - i S r ^ ^ ^ i Îi2r.*l|* t| 2 ~ - a S S " J s g *S2 _ f _ r 2 o O v - T " 5 1 ' O N O.OVc ^ s u 1 3 _ u o 02 03 CQ- > ' c t > ' o J = Co j j ^ ^ - ^ c d ^ a * ^ . w 3 o-'o . U -o 13 c3" r t ^ ' J3^ tH X - ^ r -= 1fc3-gD. g.S w o

S S3'g - - " g-S S S .H, . - a C T 3 o S t i " g - c gI ^P.H! -sjN . ^Si--gô1|l-s-s^ll-s-s8 |||-S J ^ S i-2|8 S ' g " i | 8 g | | ' g g Sll|fll-2!ll|1-vl ffl|liisi?lI|i|iS:iillilililllllllllllilllllllllllllillls o r H o o

.*) ON g- r c oo * - 5S|.S oo l g . - - 2 S S -Jj * .0 2 S " - r^ g * 2 -3 S 2 ^ _r -g Sg vi _: Ki ^ ~^ -*^ cd L ,ir *- s j< ." * S , K- ts 2 5 2 tsg . ~ H S M .. c S3 c tS * oos . .goo-^-aâ -^s S3 " St "1 ^ > ^ - ^ n u" B - Q J - g - c N ] Q ^ - 3 - g gX Q


13/43

RISK AN D PROTECTIVE FACTORS FO R DRUG PROBLEMS 75s i *1 1ao o.f 111|| o8 ?! 2S K g

SiS - M ^ - O

JlfPjimsE r : U

>.3to

"rtOao"oUH

O


15/43

RISK AN D PROTECTIVE FACTORS FO R DRUG PROBLEMS 77

inu

eewosoa

at

SCAiapee

oc

00c

urev

smonfogo

(-1

lasow

tannncmp

t/i

uancc

ct/5uI*-3

jcanus

& shwbc lanhog

penod

ooac

mauu

1II

00 00 -tJ "^CT\ - ^ ' o i-T > ~

. r. - -.^ O *0 ,O

3 % C.B -ill1) fll ^ '_ M cd *

1^1a 3 " >,o ^S o g>" ga K faa a iso I| l.. - e

vo -_;_;.-:^i E I

cdK

ONo\"3uoS oSCO

O 00S g< T3


16/43


^IKeaF

C0c:uc:&c""^

soskaoo

a

nn

o y^ 3W

o3^

C0C3.9~B.S

ac3atC81'>w

T33ia_o3w

|s

llSllllllilillllllilljtiiHIIllllI$ 3^

s c '>s* O* oa ooc -S o J


17/43

RISK AND PROTECTIVE FACTORS FOR DRUG PROBLEMS 79

"Q tflif-l3 ON .^ ON >ooif, - 2 ON^ oo 3 "~12 2 S l 3Ci

_-s sa*.sa S6 2C ONU .C *.C ^ca ~

^

_u2o

o EV^ ^o. E11

.i gO 3f K ?2 S" S ' su j- 3

. 3 00

Jin mi

. j j< a


18/43

80 D . H A W K I N S , R. CATALANQ AND J. M I L L E R

'"^Scoit-V1F

co&c0oJj

oll_^ CC!ll,0tn -^t2 4*& %w

>,a3Sri

UlCO1"ag

^4tfs(5

1Ico1>*

gcsu

_o3w

i_tot32

i~ edu >-10 o 2 *3_. o c-a a ^ fc t HW> ft>. 3 s111 I5 M O S|2 g

&3OD "Og g.s ea 3 (-S w K .S

coS J2> gi|3 a- ." -t> > 'Co> o X! _C2: 2 t> -SQ

>ndnn

jaeo

:sgeea

;eomoe

Z. - 3o e^ -j- U U fiO S k " e *" S S '!/>'oS 3 gS r 2?-.i 2 o sss S

ao3S1tT3

*/%OOON^418

i11^^11o.

1 8 8 8 * o-S

-

w

US


19/43

RISK AN D PROTECTIVE FACTORS FOR DRUG PROBLEMS 81different ethnic groupsfor example, in association with dif-ferences in the extent to which members find consumptionsocially acceptable (Flasher & Maisto, 1984; Vaillant, 1983).The second view of the law's effect focuses on supply anddemand. As noted above, legal restrictions that influence theavailability or price of alcohol or other drugs, such as taxationor laws regarding sales, appear to limit consumption.

Legal restrictions on the purchase of alcohol and norm s unfa-vorable toward alcohol use clearly are associated with a lowerprevalence of alcohol abuse. Conversely, laws and norms thatexpress greater tolerance for the use of alcohol are associatedwith a greater prevalence of alcohol abuse. Johnston (1991) hassuggested a similar relationship between n orms regarding ille-gal drugs and the prevalence of illegal drug abuse.2. Availability. The availability of drugs is dependent inpart on the laws and norm s of society. Nevertheless, ava ilabil ityis a separable factor. Wh ether or not particular substances arelegal, their availability may vary and is associated with use.Research has shown that when alcohol is more available, theprevalence of drinking, the amount of alcohol consumed, andthe heavy use of alcohol all increase (Gorsuch & Butler, 1976).With regard to illegal drugs, Maddahian, Newcomb, andBentler (1988) found in an adolescent sample that two measuresof drug availability were significantly related to the use of ciga-rettes, alcohol, marijuana, and other illegal drugs, even aftercontrolling for the am ount of money available to the subjects.Dembo, Farrow, Schmeidler, and Burgos (1979) reported thatthe availability of drugs affected substance use indirectlyamong junior high school youths. G. D. Gottfredson (1988)found that dru g availability varied in different schools and thatdrug availability influenced the use of drugs beyond the influ-ence of individual characteristics of subjects.3. Extreme economic deprivation. Indicators of socioeco-nomic disadvantage, such as poverty, overcrowding, and poorhousing, have been shown to be associated with an increasedrisk of childhood conduct problems and delinquency (Bursik &Webb, 1982; Farrington et al., 1990). However, research on so-cial class and drug use has not always confirmed popular stereo-types. A slight positive correlation between parental educationand high school seniors' marijuana use has been reported(Bachman, Lloyd,& O'Malley,1981). R. A. Zucker and Harford(1983) found that parental occupational prestige and educationwere positively related to teenage drinking. D. M. Murray,Richards, Luepker, and Johnson (1987) found that mother'soccupation was positively correlated wit h month ly alcohol use,heavy alcohol use, and marijuana use amon g seventh-grade stu-dents. The 1988 National Household Survey on Drug Abuserevealed significantly higher lifetime prevalence rates for mari-juana use among those with some college education as com-pared with those who had less than a high school education(Adams et al., 1990). In contrast, Robins and Ratcliff (1979)found that extreme poverty, though no t lower-class status pe rse, was one of three factors that increased the risk of adultantisocial behavior, including alcoholism and illegal drug use,among children wh o were highly antisocial in childhood.In sum ma ry, wherea s there ap pears to be a negative relation-ship between socioeconomic status and delinquency, a similarrelationship has not been found for the use of drugs by adoles-cents. Only when poverty is extreme and occurs in conjunction

with childhood behavior problems has it been shown to in-crease risk for later alcoholism and drug problems.4. Neighborhood disorganization. Neighborhoods w ith highpopulation density, lack of natural surveillance of public places(C. A. Murray, 1983), high resid ential mobility, physical deterio-ration, low levels of attachment to neighborhood (Herting &Guest, 1985), and high rates of adult crime also have high ratesof juvenile crime (Wilson & Herrnstein, 1985) and illegal drugtrafficking (Fagan, 1988). Simcha-Fagan and Schwartz (1986)assessed the contextual effects of neighborhood on delinquencyand found that community economic level and commun ity dis-order-criminal subculture were significantly related to offi-cially recordeddelinquency.When neighborhoods undergo rapid population changes,victimization rates increase, even after accounting for race andage differences (Sampson, 1986; Sampson, Castellano, & Laub,1981). Neighborhood disorganization has been hypo thesized tocontribute to a deterioration in the ability of families to trans-mit prosocial values to children (W McCord & McC ord, 1959;Reiss, 1986; Shaw & McKay, 1969). Although few studies ofneighborhood disorganization have explicitly exam ined its re-lationship with drug abuse, a deterioration in parenta l sociali-zation and supervision associated with neighborhood disor-ganization could also be expected to produce high rates of druginvolvement. More research is required to de term ine the effectsof neighborhood disorganization on adolescent drug abuse.

Individual and Interpersonal FactorsCertain characteristics of indiv iduals and of their personalenvironments are associated with a greater risk of adolescentdrug abuse. These characteristics are summarized below asRisk Factors 5 through 17.5. Physiological factors. Sensation seeking and low harmavoidance predict early-onset alcoholism (Cloninger et al.,1988). Poor impulse control in childhood predicts frequentmarijuan a use at age 18 (Shedler & Block, 1990). Zuckerman(1987) has suggested that sensation seeking is linked biochem i-cally to platelet monoam ine oxidase (MAO)activity, which hasalso been found to be associated with early-onset alcoholism(Tabakoff& Hoffman, 1988; von Knorring, O reland, & vonKnorring, 1987).The enzym e aldehyde dehydrogenase (ALDH), important inthe decomposition of ethanol in the body (Li, 1977), has alsobeen linked to alcoholism. Asians without one ALDH enzymedrink less and have lower rates of alcoholism than controls(Harada, Agarwal, Goedde, & Ishikaw a, 1983; Schuckit,1987;Suwaki & Ohara, 1985).Researchers have also studie d differences in genetically me -diated biological responses to alcohol amon g children of alco-holics and nonalcoholics. Pollock, Volavka, and Goodwin(1983) reported more slow-wave activity on the electroenceph a-logram (EEG) for children of alcoholics compared with chil-dren of nonalcoholics. Schuckit, Parker, and Rossman (1983)found differences in children of alcoholics and children of non-alcoholics in serum prolactin response to administration of al-cohol. Schuckit (1980) reported greater muscle relaxation inresponse to ethanol, and Schuckit and Rayes (1979) found in-


20/43

82 D. HAWKINS, R. CATALANO, AND J. MILLERcreased levels of acetaldehyde after administration of alcohol insons of alcoholics when compared with sons of nonalcoholics.

Researchers have sought to assess the independent contribu-tion of genetic factors to the development of alcoholismthrough twin and adoption studies. Kaij (1960) and Hrubecand Omenn (1981) found that among males, monozygotictwins were more than twice as likely as dizygotic twins to beconcordant for alcoholism, although in a study of both femalesand males, Curling, Clifford, and Murray (1981) reported con-cordance rates fo r alcoholism of 21 % in monozygotic and 25%in dizygotic twins.

Adoption studies in Denmark, Sweden, and the UnitedStates have provided more consistent evidence for genetictransmission of alcoholism in males, reporting rates of alcohol-ism ranging from 18% to 27% for the adopted sons of alcoholicscompared with only 5% to 6% for adopted males without abiological alcoholic parent (Bohman, 1978; Cadoret, Cain, &Grove, 1980; Cadoret & Gath, 1978; Goodwin et al., 1974;Goodwin, Schulsinger, Moller, Mednick, & Guze, 1977). Noconsistent evidence for genetic transmission of alcoholism infemales has been reported (R. M. Murray & Stabenau, 1982).

Note that the adoption studies suggesting a genetic factor inmale alcoholism also reveal that less than 30% of the sons ofalcoholics themselves become alcoholic. Furthermore, abouthalf of hospitalized alcoholics do not have a family history ofalcoholism (Goodwin, 1985), suggesting that factors other thangenetic predisposition also contribute to alcoholism.It is beyond the scope of this article to review thoroughly therecent developments in this area of alcoholism studies. Re-search continues to point toward differences in physiologicalresponses to ethanol among sons of alcoholics (Schuckit, 1987)and to other possible genetic and biochemical "markers" of riskfor alcoholism (K . Blum et al., 1990; TabakofF & Hoffman,1988). Early-onset alcoholism that is associated with impulsi-vity and aggression apparently has a partial foundation in indi-vidual physiological characteristics.Little research has been conducted on genetic predispositionand the abuse of drugs other than alcohol in humans, althoughthere is evidence from animal studies of a heritability in predis-position to barbiturate and morphine abuse (Marley, Miner,Wehner, & Collins, 1986).

6. Family alcoholand drug behaviorand attitudes. Familiesaffect children's drug use behaviors in a number of ways.Beyond the genetic transmission of a propensity to alcoholismin males, family modeling of drug using behavior and parentalattitudes toward children's drug use are family influences re-lated specifically to the risk of alcohol and other drug abuse.Poor parenting practices, high levels of conflict in the family,and a low degree of bonding between children and parentsappear to increase risk for adolescent problem behaviors gener-ally, including the abuse of alcohol and other drugs (Brook,Brook, Gordon, Whiteman, & Cohen, 1990). In this section,the family risk factors specific to alcohol and other drug abuseare reviewed. Family factors more generally predictive ofadoles-cent problem behaviors are reviewed in subsequent sections.

Parental and sibling alcoholism (Cloninger, Bohman, Sig-vardsson, & von Knorring, 1985; Cotton, 1979; Goodwin,1985) and illegal drug use (G. M. Johnson, Schoutz, & Locke,1984) increase the risk of alcoholism and drug abuse in chil-

dren. Parental drug use is associated with initiation of use byadolescents (G. M. Johnson et al., 1984; Kandel, Kessler, &Margulies, 1978; McDermott, 1984) an d with frequency ofmari juana use (Brook et al., 1990). Similar findings have beenreported fo r adolescent drinking habits (Rachal et al., 1982;R. A. Zucker, 1979). G. M. Johnson et al. (1984) found thatparental use of mari juana was associated with adolescents' useof other illegal drugs, including cocaine and barbiturates.Ahmed, Bush, Davidson, and lannotti (1984) examined theeffects of parental modeling of drug use on children's expecta-tions to use drugs and on their drug use. In a study of 420children in grades kindergarten (K) to 6, they found "salience,"a measure of the number of household users of a drug and thedegree of children's involvement in parental drug-taking behav-ior, to be the best predictor of both expectations to use andactual use of alcohol. Salience was also a predictor of children'scigarette and marijuana use. The importance of number ofhousehold users varied across substance. As the number of fam-ily members who used alcohol or marijuana increased, so didthe probability that the child used or expected to use thesesubstances. For cigarette smoking, having one householdmember who smoked cigarettes almost doubled the probabilitythat a child smoked or expected to smoke, but additionalsmokers in the home did not increase this probabil ity further.

Note that Hansen et al.'s (1987) structural equation modelinganalyses using cross-sectional data indicated indirect, but notdirect, effects of parental modeling of drug use on children'sdrug use in early adolescence. Parental modeling was directlyrelated to friends' use of drugs, which, in turn, was related tosubjects' drug use. This finding is consistent with the findingsof Brook et al. (1990), whose combined longitudinal and cross-sectional studies revealed that nondrug use and emotional sta-bility in fathers enhanced th e effect of peer nonuse of drugs an dthat psychological stability in mothers offset the effects of peerdrug use.Brook, Whiteman, Gordon, and Brook (1988) examined therole of older brothers in younger brothers' drug use and foundthat older brothers' advocacy of drugs and modeling of drug usewere both associated with younger brothers' use. They also ob-served an interaction pattern in which some of the negativeeffects of parental drug use were offset by the older brothers'nonuse. Older brothers' and peers' drug modeling both weremore strongly associated with younger brothers' use than wasparental modeling of drug use.

McDermott's (1984) research indicated that although paren-tal drug use and adolescent drug use are related, suggesting themodeling effect discussed above, permissive parental attitudestoward drug use as perceived by youths may be of equal orgreater importance than actual parental drug use in determin-ing the adolescent's use of drugs. This finding is consistent withHansen et al.'s (1987) results. Similarly, Barnes and Welte (1986)found that parental approval of drinking was a significant pre-dictor of the amount of alcohol consumed by teenage drinkers,and Brook, Gordon, Whiteman, and Cohen (1986) found thatparental tolerance of drug use predicted adolescent drug use.This relationship has been shown for Whites, Hispanics, Afri-can Americans, Native Americans and Asian Americans (Jes-sor, Donovan, & Windmer,1980).7. Poor an d inconsistent family management practices.


21/43

RISK AND PROTECTIVE FACTORS FOR DRUG PROBLEMS 83Kandel and Andrews (1987) found that lack of maternal involve-ment in activities with children; lack of, or inconsistent, paren-tal discipline (see also Baumrind, 1983; Penning & Barnes,1982); and low parental educational aspirations for their chil-dren predict initiation of drug use. Stanton (1979), Kaufmanand Ka ufm an (1979), and Ziegler-Driscoll (1979) suggestedthat familial risk factors include a pattern of overinvolvementby one parent and distance or permissiveness by the other.Differences between the effects of mothers' and fathers' dis-ciplinary techniques were observed by Brook et al. (1990). Ma-ternal control techniques were more important than paternaltechniques in explaining adolescent marijuana use. Specifi-cally, mothers' control patterns that included setting clear re-quirem ents for responsible behavior led to less marijuana use,and mothers' use of guilt to control was correlated with greaterdrug use.Baumrind (1983) classified parenting styles as authoritative,autho ritarian, or permissive and found that children who werehighly prosocial and assertive generally came from authorita-tive families. She found that parental nondirectiveness or per-missiveness contributed to hig her levels of drug use. Reilly(1979) found that common characteristics of families with ado-lescent dru g abusers included negative comm unication pat-terns (criticism, blaming, lack of praise), inconsistent and un-clear behavioral limits, and unrealistic parental expectations ofchildren.Shedler and Block (1990) found that the quality of mothers'interaction with their children at age 5 distinguished childrenwho were frequent users of marijuana at age 18 from those whohad only experimented with marijuana use. Mothers of chil-dren who became frequent users were relatively cold, underre-sponsive, and underprotective with their children , giving theirchildren little encouragement bu t pressuring them to performin tasks.

Norem-Hebeisen, Johnson, Anderson, and Johnson (1984)also found that the quality of adolescents' relationships withtheir parents was related to patterns of drug use. Generally,drug users perceived the ir fathe rs as more hostile, rejecting , andadversarial th an did nonusers.The evidence suggests an independent contribution of familyinteractions to adolescent drug use, separate from the effects ofparental drug use. Tec (1974) found that parental drug use in arewarding family structure only slightly promoted frequentmarijuana use but that in a unrewarding context, there was aclear association between levels of drug use by parents and theirchildren.In sum ma ry, the risk of drug abuse appears to be increasedby family management practices characterized by unclear ex-pectations for behavior, poor monitoring of behavior, few andinconsistent rewards for positive behavior, and excessively se-vere and inconsistent punishm ent for unwanted behavior.8. Family conflict. Although children from homes brokenby marita l discord are at higher risk of delinquency and druguse (Bau mrin d, 1983; Penning & Barnes, 1982; Robins, 1980),there does not appear to be a direct independent contributionof "broken homes" to delinquent behavior (Wilson & Herrn-stein, 1985). Conflict among family members appears moreimportant in the prediction of delinquency than does familystructure per se (Farrington, Gallagher, Morley, Ledger, &

West, 1985; McCord, 1979; Rutter & Ciller, 1983). Rutter andGiller have noted th at parental conflict is associated with anti-social behavior in children even when the home is unbroken(see also W McCord & McCord, 1959; Porter & O'Leary, 1980)and that even in samples in which all homes are broken, theextent of family conflict is associated with the likelihood ofantisocial behavior in the children (see also Hetherington, Cox,& Cox, 1979; Wallerstein & Kelly, 1980). Similarly, Simcha-Fa-gan, Gersten, and Langner (1986) found that the use of heroinand other illegal drugs was strongly associated with parentalmarital discord. In summary, children raised in families highin conflict appear at risk for both delinquency and illegaldrug use.

9. Low bonding to family. Parent-child interactions charac-terized by lack of closeness (Brook , Lukoff, & Whiteman, 1980;Kandel et al., 1978) and lack of maternal involvement in activi-ties with children (Braucht, Kirby, & Berry, 1978; Penning &Barnes, 1982) appear to be related to initiation of drug use.Conversely, positive family relationshipsinvolvement and at-tachmentappear to discourage youths' initiation into druguse (Brook et al., 1986; Gorsuch & Butler, 1976; Jessor & Jessor,1977; Kim, 1979; Norem-Hebeisen et al., 1984; Selnow, 1987).Hundleby and Mercer (1987) found that adolescents' reports ofparental trust, warmth, and involvement explained small por-tions of the variance in the extent of tobacco, alcohol, and mar i-juana use.Bonding to family may inhibit drug involvement durin g ado-lescence in a ma nne r similar to the w ay in which family bond-ing inhibits delinquency (Hirschi, 1969). Brook et al. (1990)pointed to the salience of parent-child attachment in describ-ing the pathways to marijuana use frequency in their combinedlongitudinal and cross-sectional studies. They reported acausal pathway in which parental internalization of traditionalvalues led to the development of strong parent-child attach-ment; this mutual attachment led to the child's internalizationof traditional norms and behavior, whic h in turn led the young-ster to associate with non-drug-using peers, which led tononuse.

10. Early and persistent problem behaviors. The greater thevariety, frequency, and seriousness of childhood antisocial be-havior, the more likely antisocial behavior is to persist intoadulthood (Robins, 1978).A longitudinal study of 5-year-olds followed into adulthood(Lerner & Vicary, 1984) found that a difficult temperament,including frequent negative mood states and withdrawal, con-tributes to drug problems. Children characterized by with-drawal responses to new stim uli, biological irregularity, slowadaptability to change, frequent negative mood expressions,and high intensity of positive and negative expressions of affectmore often became regular users of alcohol, tobacco, and ma ri-juana in adulthood than "easy" children, who evidencedgreater adaptability and positive affect early in life. Similarly,Shedler and Block (1990) found that frequent marijuana usersat age 18 were characterized in childhood by emotional dis-tress. Lerner and Vicary (1984) suggested that the negativemood and withdra wal responses of the difficult child may beanalogous to the depression and social alienation frequentlyreported for drug abusers (Knight, Sheposh, & Bryson, 1974;


22/43

84 D. HAWKINS, R. CATALANO, AND J. MILLERPaton & Kandel, 1978; Paton, Kessler, & Kandel, 1977; Smith &Fogg, 1978).Brook et al. (1990) found that children who were irritable,easily distractible, had temper tantrums, fought often with sib-lings, and engaged in predelinquent behavior were more likelyto use drugs in adolescence.Aggressive behavior in boys appears to signal another pathtoward later antisoc ial behav ior. Aggressiveness in boys as earlyas ages 5-7 (Grades K-2) has been found to predict later antiso-cial behavior includ ing frequent drug use in adolescence (Kel-lam & Brown, 1982), drug problems in adulthood (Lewis,Robins, & Rice, 1985; Nylander, 1979), and delinquency in ado-lescence (Loeber, 1988; Spivack, 1983). However, early aggres-siveness is not invariably followed by serious antisocial behav-ior. App roximately 30% to 40% of the boys engaged in maladap-tive, aggressive behaviors continue that behavior 4 to 9 yearslater (Loeber & Dishion, 1983).Few youths develop highly physically aggressive behav iors inlate child hood or adolescence if not engaged in such behaviorsin earlier childhood, and most boys grow out of early aggressivebehaviors. However, if aggressive behavior continues into earlyadolescence (age 13), it is a relatively strong predictor of contin-ued aggressive behavio r in late adolescence as well as of lateralcoholism (Loeber, 1988; McCord, 1981). Furthermore, if anti-social behavior persists and becomes more varied in early ado-lescence to include fighting and school misbehavior, drug abuseis more likely (Barnes & Welte, 1986; Kandel, 1982). Barnes andWelte found that school misconduct was one of the three mostimportant predictors of alcohol-related problems in a study ofsubjects from six ethnic groups in Grades 7-12.Hyperactivity and attention-deficit disorders have beenshown to increase risk for delinque ncy when combined withconduct problems incl udin g aggression (Loney, Kram er, & Mi-lich, 1979). Gittelman, Mannuzza, and Bonagura (1985) founda higher prevalence of substance abuse disorders in late adoles-cence among subjects diagnosed as hyperactive in childhood.As with delinqu ency, those at highest risk were those with bothhyperactivity and con duct disorders. The G ittelman et al. find-ing that hyperactivity, without accompa nying conduct prob-lems, predicts an increased risk of substance abuse has not beenreplicated. However, it suggests further investigation into therelationship betwee n attention-deficit disorders, conduct prob-lems, and substance abuse.11. Academic failure. Although there is an inverse relation-ship between intellectual ability and delinquency after control-ling for socioeconomic status and race (G. D. Gottfredson,1981), a simila r relationship has not been reported for drug use,in spite of the covariation in delinquent and drug-using behav-iors. In fact, in an African-American inner-city sample, higherscores on reading readiness and IQ tests in Grade 1 predictedearlier and more frequent use of alcohol in adolescence (Flem-ing, Kellam, & Brown, 1982). Similarly, in a national probabil-ity sample, hig h intelligence, as assessed by the Armed ForcesQualifying Test, was associated with higher lifetime levels ofcocaine use amon g you ng adults age 19-26 (Kandel & Davies,1991).Nevertheless, failure in school has been identified as a pre-dictor of adolescent drug abuse (Jessor, 1976; Robins, 1980).Poor school performa nce has been found to predict frequency

and levels of use of illegal drugs (Smith & Fogg, 1978). Holm-berg (1985), in a longitudinal study of 15-year-olds, reportedthat truancy, placement in a special class, an d early drop ou tfrom school w ere prognostic factors for drug abuse. 1 n contrast,outstanding performance in school reduced the likelihood offrequent drug us e among a ninth-grade sample studied byHundleby and Mercer (1987).What is not clear from the existin g research is when, develop-mentally, poor school achievement becomes a stable predictorof drug abuse. The available evidence suggests that social ad-justment is more importan t than academic performance in theearly elementary grades in predicting later drug abuse. Earlyantisocial behavior in school may predict both academic failurein later grades (Feldhusen, Thurston, & Benning, 1973) andlater drug abuse. Academic failure in late elementary gradesma y exacerbate the effects of early antisocial behavior or con-tribute independently to dru g abuse.12. Low degree of commitment to school. A low degree ofcommitment to education also appears to be related to adoles-cent drug use. A nnu al surveys of high school seniors by John-ston, O'Malley, and Bachm an (1985) show that the use of hallu-cinogens, cocaine, heroin, stimulants, sedatives, or nonmedi-cally prescribed tranquilizers is significantly lower amongstudents wh o expect to attend college than am ong those who donot plan to go on to college. G. D. Gottfredson (1988) found thattruancy for both boys and girls wa s associated with drug involve-ment, after accounting for effects of ethnicity, parental educa-tion, and delinquency. Factors such as how much students li keschool (Kelly & Balch, 1971), time spent on homework, andperception of the relevance of course work are also related tolevels of drug use (Friedman, 1983), indicating a negative rela-tionship between comm itment to education and frequent druguse among junior and senior high school students.13. Peer rejection in elementary grades. Although it would

be premature to posit a direct link between peer rejection andsubstance abuse, low acceptance by peers seems to put an ado-lescent at risk for school problems and crim inali ty (Coie, 1990;Kupersmidt, Coie, & Dodge, 1990; Parker & Asher, 1987),which are also risk factors for substance abuse (Hawkins,Lishner, Jenson, & Catalano, 1987).Little research has been done on the direct link between peerrejection and substance use, but traits of the child that havebeen associated with peer rejectionaggressiveness, shyness,and withdrawalhave been examined for their relationship todrug use. For example, Kellam, Ensminger, and Simon (1980)found that children who had been shy in first grade reportedlow levels of involvement in drug use, whereas those who hadbeen aggressive or had shown a co mbination of aggressivenesswith shyness in first grade had the highest levels of use. Brook etal. (1986) found that c hildh ood traits relevant to peer rejectionsocial inhibition, isolation from peers, and aggressionagainst peerswere not significantly associated with adoles-cent drug use stage. However, aggression against peers duri ngadolescence was associated with stage of use, and teenagerswho were less socially inhibited and less isolated from peerswere likely to be at a m ore advanced stage of use.These studies suggest that the link between peer rejectionand subsequent drug use may not be a simple one. Shyness, byisolating a child from his or her peers, ma y protect the child


23/43

RISK AND PROTECTIVE FACTORS FOR DRUG PROBLEMS 85against drug use by elim inating one source of influence to use:drug-using peers. Aggressiveness, on the other hand, thoughresulting for some children in exclusion from groups of conven-tional peers, may be associated with acceptance by other ag-gressive and perhaps delinquent peers who could foster druguse (Cairns, Cairns, Neckerman, Gest, & Gairepy, 1988).Ha rtup (1983) suggested th at rejected children form friend-ships with other rejected children during the preadolescentyears and that these friendship groups become d elinquent dur-ing adolescence. However, this process is as yet unconfirmed(Tremblay, 1988).

14. Association with drug-using peers. Peer use of sub-stances has consistently been found to be among the strongestpredictors of substance use among youth (Barnes & Welte,1986; Brook et al., 1990; Elliott, Huizinga, & Ageton, 1985;Jessor et al., 1980; Kandel, 1978, 1986; Kandel & Andrews,1987). Studies among specific ethnic groups c onfirm th is rela-tionship. New comb and B entler (1986) reported that the influ-ence of peers on adolescent drug use was stronger than that ofparents for Whites, African Americans, Asian Am ericans andHispanic Am ericans. Similar findings were reported by Byramand Fly (1984). Harford (1985) found that African-Americanyouths who did not dr ink alcohol reported fewer school friendswho drank than did those who drank, and Dembo et al. (1979)found that frien ds' use of alcohol and m ariju ana was related toa youth's own use for both A frican-A meric an and P uerto Ri-can-American youths.75. Alienation and rebelliousness. Alienation from the dom-inant values of society (Jessor & Jessor, 1977; Kandel, 1982;Penning & Barnes, 1982), low religiosity (Jessor et al., 1980;Kandel, 1982; Robins, 1980), and rebelliousness (Bachman etal., 1981; Kandel, 1982) have been shown to be positively relatedto dru g use and delinqu ent behavior. Shedler and Block (1990)found that in terpersona l alienation measured at age 7 predictedfrequent marijuana use at age 18. Similarly, high tolerance ofdeviance (Jessor & Jessor, 1977), a strong need for indepen-dence (Jessor, 1976), and n ormlessness (Paton & Kandel, 1978)have all been linked with drug use. Al l these qualities wouldappear to characterize youths who are not bonded to society.

16. Attitudes favorable to drug use. Research also has showna rela tionship between drug use initiation and specific attitudesand beliefs regarding drugs. Initiation into use of any substanceis preceded by values favorable to its use (Kandel et al, 1978;Krosnick & Judd, 1982; Smith & Fogg, 1978).17. Early onset of drug use. Early onset of drug use predictssubsequent misuse of drugs. Rachal et al. (1982) reported thatmisusers of alcohol appear to begin drinking at an earlier age

than do users. The earlier the onset of any drug use, the greaterthe involvement in other drug use (Kandel, 1982) and thegreater the frequency of use (Fleming, Kellam, & Brown, 1982).Earlier initiation into dru g use also increases the probability ofextensive and p ersistent involvement in the use of more danger-ous drugs (Kandel, 1982) and the pro bability of involvement indeviant activities such as crime and selling drugs (Brunsw ick &Boyle, 1979; O'Donnell & Clayton, 1979). Robins and Przybeck(1985) found that the onset of drug use before the age of 15 was aconsistent predictor of drug abuse in the samples they studied.Conversely, a later age of onset of drug use has been shown to

predict lower drug involvement and a greater probability ofdiscontinuation of use (K andel, Single, & Kessler, 1976).Implications of Research on Risk

A risk-focused prevention approach requires identificationof those risk factors to be addressed. Unfortunately, all the in-formation needed to select the most promising risk factors forintervention is not yet at hand. Experimental research isneeded to discover which risk factors are causal and w hich arespurious in the etiology of drug abuse. Only by addressing riskfactors in experimental trials and observing the effects on drugabuse can one determine whether a precursor of drug abuse iscausally related to drug abuse. Experimental prevention re-search is therefore necessary both to unde rstand the etiology ofdrug abuse and to determine which risk factors should be tar-geted in preve ntion policy and programs.Several general conclusions regarding risks for drug abusecan be draw n, which have implications for prevention. First,the risk factors reviewed above have been shown to be stableover time in spite of changing norms. For example, despitegeneral changes in norms regarding the use of drugs such asmarijuana over the past 20 years (Johnston, O'Malley, & Bach-man, 1989), studies conducted in different times and placeshave shown these factors to predict adolescent d rug abuse rela-tively consistently. This suggests the risk factors' stability aspredictors and their viability as targets for preventive work .Second, risk factors from several domains predict drugabuse. Some factors are characteristics of the individ ual; othersare characteristics of families an d their interactions, schoolsand classroom experiences, peer groups, and broader commu-nity, legal, economic, and cultural factors.Third, different risk factors are salient at different periods ofdevelopment. For example, poor academic achievement inGrades 1 and 2 does not appea r to be a stable predictor ofteenage drug abuse (Kellam & Brown, 1982), though poorachievement in the later grades predicts drug abuse. Aggressive-ness at ages 5-7 predicts later drug abuse and, if it continues,becomes more strongly predictive of drug abuse w ith increas-ing age.Fourth, there is evidence that the more risk factors present,the greater the risk of drug abuse (Bry, McKeon, & Pandina,1982; Newcomb et al, 1986). Ru tter (1980) found a multiplica-tive effect of added risk factors on the likeliho od of childhoodpsychopathology, and Newcomb, Maddahian, Skager, andBentler (1987) reported a similar contribution of combinationsof different risk factors to overall risk for adolescent drug use. Itis plausible that a greater length of exposure to environmentalrisk factors exacerbates risk as well. Curre nt research focuses onhow risk factors interact in the etiology of drug abuse (Loeber &Stouthamer-Loeber, 1986). Greater precision in estim ating howmuch various risk factors contribute to dru g abuse w ill help tofocus prevention efforts on those risk factors that are most viru-lent.A risk-focused prevention approach does not require that riskfactors be m anipula ted directly. It may be impossible to reduceor chang e certain risk factors directly through preventive inter-vention. In these instances, the goal of prevention efforts willbe to mediate or moderate the effects of the identified bu t non-


24/43

86 D. HAWKINS, R. CATALANO, AND 1 MILLERmanipulable risk factors. A family history of alcoholism, forexample, may be difficult or impossible to change. Neverthe-less, it may be possible to moderate the effects of a family his-tory of alcoholism by interveningwith children who are at riskbecause of their exposure to this environment. One task ofrisk-focused prevention research is to determine which risk fac-tors can be manipulated, which risk factors cannot be changedbut can be mediated or moderated, and which risk factors can-not be affected at all.

Protective Factors Against Drug AbuseBecause some risk factors for drug abuse may be resistant orimpossible to change, the results of research on protective fac-tors are important for prevention policy. Protective factors medi-ate or moderate the effects of exposure to risk (Cowen & Work,1988; Garmezy, 1985; Rutter, 1985; Werner, 1989). To the extentthat protective factors are identified that inhibit drug abuseamong those at risk, strategies can seek to address risk by en-hancin g these protective factors. Research with populations ex-posed to multiple risks has identified substantial subgroups ofindividuals who are able to negotiate risk exposure successfully,escaping relatively unscathed (Werner, 1989). These observa-tions have led to interest in the etiological importance of factorsthat may protect against health problems including drug abuse.Concepts of vulnerability and resiliency have been advancedto identify the extent of individual susceptibility to risk (Rutter,1985). Vulne rability denotes intensified susceptibility to risk;resiliency is the ability to withstand or surmount risk. Fromthis perspective, protection involves enhancing resilient re-sponses to risk exposure. The hypothesis is that certain charac-teristics or conditions mediate or moderate the effects of expo-sure to risk, thereby reducing the vulnerability and enhan cingthe resiliency of those at risk an d protecting them from unde-

sirable outcomes. To illustrate, Werner and Sm ith (1982) foundthat in rural K auai, Haw aii, being raised in a small family withlow conflict, having high intelligence, and being a firstbornchild buffered the effects of extreme poverty and other riskfactors for poor educational, economic, and health outcomes.For the concept of protective factorsas distinct from riskfactorsto be useful, it must apply to differences in outcomesamong ind ivid ual s exposed to the same risks. Though som ehave viewed protective factors simply as the opposite of thosevariables identified as risk factors (Labouvie & McGee, 1986),this conception does not appear particularly useful. Designat-ing two distinct constructs (e.g., risk and protective factors) todistinguish extreme levels of a single variable bearing a line arrelationship to drug abuse adds little. It is not necessary topostulate protective factors if better outcomes are observed inthose not exposed to risk. On the other hand, if protective fac-tors are viewed as sources of differences in response to a givenamount of exposure to risk, the construct stimulates attentionto nonlinear and interactive relationships among risk and pro-tective factors.In urging a focus on protective mechanisms, Rutter (1985)described interactive processes to identify multiplicative inter-actions or synergistic effects, in which one variable potentiatesthe effect of another. The idea of identifying protective pro-cesses or specifying particular interactions among variables

that produce an endu ring shield or resilience in the face of riskfor negative outcomes has direct relevance for risk-focused drugabuse prevention. It suggests that the goals of risk-focused pre-vention may be accomplished both through direct efforts atrisk reduction an d through the enhancement of protective fac-tors that moderate or mediate the effects of exposure to risk.Preventive work tha t seeks to address risk factors for drug abusemust clearly hypothesize how a particular intervention is ex-pected to address risk: by directly eliminating or reducing a riskfactor or by mediating or moderating its effects through theenhancement of protective factors or processes.Little research has focused specifically on p rotection againstadolescent dru g abuse defined in this way. However, recentlyBrook et al. (1990) identified two mechanisms by whic h protec-tive factors reduce risk for adolescent drug use. The first is a"risk/protective" mech anism th roug h which exposure to riskfactors is moderated by the presence of protective factors. Theyreported that the risk posed by drug-u sing peers was moderatedby a strong attachment or bond between parent and adolescentand by parent conventionality. The second is a "protective/pro-tective" mech anism th roug h which one protective factor poten-tiates another protective factor, strengthening its effect. Theyreported that a strong bond of attachment between adolescentan d father enhanced the effects of other protective facto rs suchas adolescent conventionality, positive maternal c haracteris tics,an d marital harmony in preventing drug use.In related research areas, Garme zy (1985) has identified pro-tective factors among children exposed to extreme stress be-cause of highly disturbed family circumstances. These includea child's own positive temperament or disposition, a supportivefamily mi lieu , and an external support system that encouragesan d reinforces the child's coping efforts an d strengthens themby incu lcatin g positive values. Rutter (1985) has suggested thatresilient children display a repertoire of social problem solvingskills and belief in their own self-efficacy.In designing interventions to reduce the negative effects ofidentified risk factors, it is important to focus attention on thepotential positive effects of such protective factors. The avail-able evidence suggests that to be viable, a prevention strategyrequires attention to risk and protective factors related to in di-vidual vulnerability, poor child rearing, school achievement,social influences, social skills, and broad social norms, all ofwhich are implicated in the development of adolescent drugabuse. Because risks are present in several social dom ains andcumulate in predicting drug abuse, multicomponent preven-tion strategies focused on reducing multiple risks and enhanc-ing multiple protective factors hold promise. Such strategieswould be designed to build up protection while reducing risk.

Each risk factor targeted should be addressed duri ng the de-velopmental period at which it begins to stabilize as a predi ctorof subsequent drug abuse. Interventions must also target popu-lations at greatest riskgroups and individuals who are ex-posed to a large number of risk factorsif the prevalence ofdrug abuse, as defined here, is to be reduced through preven-tion efforts. Although intervention with people who are notexposed to m ultiple risk factors may delay or prevent the onsetof drug use in the general population, a desirable goal in its ownright, it may fail to reduce significantly the prevalence of drugabuse.


25/43

RISK AND PROTECTIVE FACTORS FOR DRUG PROBLEMS 87The evidence suggests a developmentally adjusted, multiple-component risk-reduction strategy that cuts across traditionalhealth, education, and human service delivery systems. Thestrategy m ust reach those at highest risk by virtu e of exposureto multiple risk factors. It must address the most significant riskfactors faced by those groups. Finally, the strategy may explic-itly seek to increase protective factors as mediators or modera-

tors against risks that cannot be changed by intervention.Using Theory to Guide PreventionResearch and Practice

To design a multicom ponent intervention strategy that seeksto reduce m ultiple risk factors and simultaneously enhanc e pro-tective factors among those exposed to risk, it is useful to beguided by a theory of causation and prevention. Theory sup-plies the explanatory framework for the observed evidence re-garding risk and protective factors for drug abuse by hypothe-sizing causal relationships among these variables that lead to-ward or away from drug abuse. Theory is also useful in guidingthe design of complementary prevention interventions in dif-ferent social un its when m ultiple interventions are desired. Toguide prevention interventions, theory should (a) identify thefactors that predict drug abuse, (b) explain the mechanismsthrough which they operate, (c) identify the factors that influ-ence these mechanisms, (d) predict points to interrupt thecourse le ading to drug abuse, and (e) specify the interventions toprevent onset of drug abuse (Ka zdin, 1990).It is not our goal to review theories of drug abuse (see Lettieri,Sayers, & Pearson, 1980, for a review). Nevertheless, an exampleillustrates how theory can provide clear direction for preventiveinterventions of the type described here.As noted by Ka zdin (1990), our delinquency and dru g abuseprevention efforts have been grounded in the social develop-ment model (Farrington & Hawkins, 1991; Hawkins & Lam,1987; Hawkins & Weis, 1985). An integration of control theory(Hirschi, 1969) and social learning theory (Bandura, 1977), thesocial development model emphasizes the role of bonding toprosocial family, school, and peers as a protection against thedevelopment of conduct problems, school misbehavior,truancy, and drug abuse. This concept of bonding is closelyrelated to the conc ept of attachm ent as denned by Bowlby(1969,1973) and as observed by Brook et al. (1990) to inhib itadolescent drug abuse. It is also consistent with Garmezy's(1985) identification of familial and external support and valuesystems as protective factors against exposure to stress in child -hood.Four elements of social bonding have been shown to be in-versely related to drug use. These are strong attachment to par-ents (Brook, Brook, et al., 1990; Brook, Gordon, et al., 1986;Hundleby & Mercer, 1987; Jessor & lessor, 1977; Norem-Hebei-sen et al., 1984); comm itment to schooling (Friedman, 1983;Johnston, Bachman, & O'Malley, 1981; Kim, 1979; Krohn &Massey, 1980); regular involvem ent in church activities (Schle-gel & Sanborn, 1979; Wechsler & McFadden, 1979); and beliefin the generalized expectations, norms, and values of society(Akers, Kro hn, Lanza-Kaduce, & Radosevich, 1979; Krohn &Massey, 1980). Research has not yet determined whe ther theseelements of social b ondin g are best viewed simply as the oppo-

site extremes of variables already identified as risk factors fordrug abuse (e.g., low commitment to schooling and alienationand rebelliousness)or whether social b ond ing represents a dis-tinct protective factor capable of buffering the effects of otherrisk factors such as a family history of alcoholism or extremepoverty. Further research on this question is needed.The social development model specifies hypotheses regard-ing the processes that produce bonding to a social unit. Interac-tions among (a) opportunities for involvement offered in eachsocial unit, (b) the skills used by individuals in these socialunits, and (c) the reinforcements offered in these units are hy-pothesized to produce social bonds of attachment, commit-me nt, and belief in the values of the social units in which youngpeople develop (see Catalano & H awkin s, 1986).Guide d by this social development perspective, our own risk-focused prevention work has two purposes: (a) to understandbetter the processesby which risk and protective factors contrib-ute to the etiology of drug abuse in adolescence and (b) to testpromising approaches to prevent adolescent drug abuse. Wehypothesize that children who develop strong bonds to socialunits holding norm s antithetical to drug abuse will be less likelyto abuse drugs.To enhance social bonding, we manipulate social settingsand individu al capacities using the principles of social learningtheory in developmentally appropriate ways. For example, inthe school setting, we train teachers in methods of proactiveclassroom ma nage men t, interactive teaching, and cooperativelearning, includin g students in peer teaching. Th e explicit,theory-driven objectives of these intervention elements are (a)to make available opportunities for children to be involved inprosocial activities, (b) to provide skills needed to undertakethese activities successfully, and (c) to provide positive reinforce-ment for successful involvement.All of these objectives servethe broader goal of strengthenin g bond ing to the social unit , inthis case the school. From a social developm ent perspective, thesame three objectives guide interventions with parents, day-care providers, youth mini sters, recreation workers, and othersparticipating in the socialization of children. The framework ofthe social development model thus fosters a multicomponentprevention approach, grounded in knowledge of risk and pro-tective factors and consistent in goals, across a va riety of socialsettings.

Current Risk-Focused Drug PreventionDuring the 1960s and 1970s there was little explicit attentionto risk or protective factors for drug abuse in the design anddevelopment of preventive interventions. More recent researchon dru g abuse prevention has focused on risk reduction, but hasno t included atten tion to m ultiple risk or protective factors and,for the most part, has not addressed risk factors that appeardevelopmentally before the age of likely drug use initiation.Most recent prevention research has targeted only two risk fac-tors for drug abuse, both of which are most salient just at thepoint of drug use initiation: (a) laws and norms favorable to druguse and (b) social influences to use drugs. Approaches targetingthese risk factors are designed for a relatively quick "return," inthat if they are effective, they should reduce or curtail drug useimmediately. Prevention approaches that target these risk fac-


26/43

D. HAWKINS, R. CATALANO, AND 1 MIL L E Rtors are summ arized below; the right half of Table 1 summa-rizes the effects of these approaches on the risk factors ad-dressed.Supply Manipulation, Interdiction, and EnforcementStrategies

Attention to the laws and norms of society related to the useof alcohol and other drugs is clearly warranted, given the linkbetween these factors and rates of alcoholism and drug abuse. Ifreduction of the prevalence of abuse of drugs is the goal, theevidence does not support those who advocate the legalizationof currently illegal drugs such as marijuana and cocaine (Clay-ton, in press). Rather, the evidence supports efforts to limitbehavior that is inconsistent with existing legal sanctions. Thishas been attempted through efforts to control the supplies ofboth legal and illegal drugs.Since the repeal of prohibition, the supply of alcohol hasbeen manipulated in several ways, including taxation, age re-strictions on consum ption, restrictions on hours of purchase,and restrictions on liquor-by-the-drink sales. As noted earlier,restricting availab ility and increasing the price of alcohol byincreasing taxes on the p urchase price can reduce rates of alco-hol abuse as indica ted in rates of cirrhosis of the liver and alco-hol-related traffic fatalities. Although increasing age restric-tions on alcohol purchases and restrictions on liquor-by-the-drink sales appears less effective than taxation in limitingalcohol abuse, these strategies also have shown desirable effectsin reducing alcohol-related traffic fatalities (Blose & Holder,1987; Decker, Graitcer, & Schafmer, 1988; Krieg, 1982).This evidence might appear to imply that supply manipu la-tion strategies such as drug interdiction and arrests of drugdealers would have a similar desirable effect on the abuse ofillegal drugs by raising the price of these drugs to the user.However, existing evidence does no t support this contention.Analysis by the Rand Corporation resulted in the conclusionthat neither a doubling of interdiction nor increased arrests ofdrug dealers would affect retail prices or the av ailability of ille-gal drugs (Polich, Ellickson, Reuter, & Kahan, 1984). Datafrom the Drug Enforcement Administration confirm this con-clusion. In spite of an increase in federal spending on interdic-tion and law enforcement from $1.807 billion in 1986 to $3.770billion in 1989, the average street price of cocaine fell from $100to $75 dollars per gram during the same period. Although atsome level well beyond current spen ding, interdiction and en-forcement might reduce drug supplies and d rive up prices, thefiscal costs, effects on U.S. international trade, and constraintson individual rights required w ould be excessive. Moreover, inthis scenario, if demand for illegal drug s were not reduced, it isplausible that domestic producers of synthetic drugs would stepin to fill demand as interdiction began to reduce drug supplies,thus continuing to hold down prices to users.In our view, the most powerful effect of interdiction and en-forcement activities is to communicate general social norms ofdisappro val for the distribu tion and use of illegal drugs. Socialnorms antithetical to use appear associated with reductions inthe prev alence of the frequent use of marijuan a (Robins, 1984)and other illegal drugs (Johnston, 1991). Supply-reduction strat-egies com mu nicate an impo rtant message to citizens but

should not be expected, by themselves, to eliminate illegal drugsupplies, to significantly raise the price of illegal drugs, or toeliminate drug abuse. Those who are at greatest risk of drugabuse by virtue of low social bonding to society ma y view therelative benefits of drug dealing and drug use as worth the risksof apprehension. The prevention of alcohol and other drugabuse among those at greatest risk requires attention to thefactors that d istingu ish these people. The risk factors encoun -tered by these persons at highest risk must be addressed toreduce the demand for illegal drugs.Changing Social Norms

A second approach curre ntly emphasized is changing socialnorms about drug- and alcohol-influenced behaviors. Th e ap-proach includes "Just Say 'NO!'" activities, community coali-tions against drugs, media campaigns, an d certain policychanges.C. A. Johnson and Solis (1983) and Perry, Klepp, and Shultz(1988) reviewed a number of comm unity health promotion pro-grams aimed at reducing cardiovascular disease by changingsmoking and other risk-related behaviors. These programs in-cluded invo lvement of the m ass media, risk-factor screeningprograms, and education programs for adults and youths. Theyhave been associated with lower smoking onset rates amongyouths (Perry, Klepp, & Shultz, 1988) and cessation or reduc-tion of smoking (C. A. Johnson & Solis, 1983).Of particu lar interest in th is area is the influence of advertis-ing on drug use. There is some indication that higher exposureto "life-style" ads prom otin g alcohol consumption is foundamong adolescents who report higher levels of drink ing (Atkin,Hocking, & Block, 1984).Th e media and advertising industries have cooperated in anational project to encourage negative attitudes toward the useof illegal drug s throu gh the use of antidru g advertising. Resultsof ma ll intercept surveys indica te that saturation advertising in10 markets was accompanied by significant normative changesover a 1-year period (Black, 1989). College students and chil-dren were more negative in their attitudes toward drugs, vieweddrug users less positively, and perceived less drug use amongtheir friends in 1988 compared with 1987. Moreover, in areasthat received saturation advertising, 9% to 15% more childrenreported conversations abou t drugs with parents, teachers, andsiblings in 1988 than did in 1987. In the balance of the UnitedStates, there were no increases in such communications. Teen-agers age 13 through 17 showed the fewest changes in attitudesin association with saturation advertising, though they becamemore positive in their views toward nonusers and perceivedgreater risks from marijuana and cocaine use (Black, 1989). Ofcourse, these differences could have been produced by otherfactors operating in communities sufficiently concerned aboutdrug abuse that their broadcast media would run a saturation-advertising campaign against drugs.Social norms regarding the use of specific drug s and theirattendant risks and benefits can change over a relatively shortperiod of time. From 1978 to 1983, the proportion of the na-tion's high school seniors who perceived there to be a greathealth risk associated with the regular use of marijuana rosefrom 38% to 63%. Over the same period, the proportion of the


27/43

RISK AN D PROTECTIVE FACTORS FO R DRUG PROBLEMS 89nation's seniors w ho used mariju ana daily dropped from 10.7%to 5.5% (Johnston, 1985).An important question for study concerns the role of broadlyfocused norm-change efforts, such as media campaigns, in pro-ducing such changes in norms and the frequent use of drugs.Studies are needed that examine how these efforts affect chil-dren at greatest risk for drug abuse. It is not k now n how chil-dren who come from poorly m anaged families, who have failedin school, who are aggressive, or who have lost commitm ent toschool respond to "Just Say NO!'" or other antidrug messagesin the media or in their personal social environments.Changes in social norms have also been codified in schoolpolicies regarding drug-using behavior (Moskowitz & Jones,1988). Recent studies of school policies regulating smokinghave shown that more comprehensive policies, which empha-size prevention of use and restrictions on opportunities for usein or near school grounds, appear to reduce the amount ofsmoking by students (Pentz, Brannon, et al., 1989), althoug heffects on smoking prevalence are less consistent. These poli-cies appear to affect smoking behavior prima rily through theclear specification of norms regarding smoking rather thanthrough the enactment of punitive consequences for policy vio-lations, which have not shown effects in reducing smoking(Pentz, Brannon, et al., 1989). Additional research is needed onthe effectiveness of school policies in preventing or reducingthe use of drugs other than tobacco and on the effects of suchpolicies on those at highest risk for drug abuse.Social Influence Resistance Strategies

As noted earlier, among the strongest correlates of teenagedrug-using behavior is association with others who use drugs. Ifthe relationship betw een association with drug-using peers anddrug-using behavior is actually causal, the manipulation of afactor that accounts for a great deal of variance in drug usewould hold promise for producing significant reductions in ado-lescent drug use.Prevention strategies focused on social influences to usedrugs also are appealing from a cost-effectiveness perspective.Because peer influence to use drugs is salient developmentallyat the point of onset of drug use, long delays are not requiredbefore effects of such interventions can be observed.The most heavily researched strategy for addressing socialinfluences to use drugs is classroom-based skills trainin g foradolescents in Grades 5 through 10,most ofte n Grades 6 and 7.The training teaches students through instruction, modeling,and role play to identify and resist influence s to use drugs and,in some cases, to prepare for associated difficulties and stressesanticipated in the process of resisting such influences (Botvin,1986). Grounded in social learning theory (Bandura, 1977),social influence resistance strategies view drug use as a sociallyacquired behavior, initiated and reinforced by drug-usingothers (Bukoski, 1986).Whereas all programs of this type offer skills in resistingsocial influences to use drugs, many also seek to promotenorms negative toward drug use (Hansen, Johnson, Flay, Gra-ham, & Sobel, 1988; Perry, 1986). These no rmativ e-chan ge com -ponents have included efforts to depict drug use as sociallyunacceptable; identification of short-term negative conse-

quences of drug use; the provision of evidence that drug use isnot as widespread among peers as children may think; encour-agement for children to make public commitments to remaindrug free; and, in some instances, the use of peer leaders toteach the curriculum (Botvin, 1986; Klepp, Halper, & Perry,1986).Social influence resistance approaches also have been com-bined with training in problem-solving and decision-makingskills, skills to increase self-control and self-efficacy, adaptivecoping strategies for relievin g stress and anxiety, interpersonalskills, and general assertive skills (Botvin, 1986; Flay, 1985). Inthis regard, Botvin's skills training program has combined ele-ments of both social in fluence resistance t rain ing and socialcompetence skills training discussed later (Botvin & Wills,1985). Recent projects have also combined classroom-based so-cial influence resistance curricula with mass-media program-ming and parent involvement strategies in com prehensive inter-ventions seeking to change norms toward drug use and increaseresistance to drug-prone influences among adolescents (Pentz,Dwyer, etal., 1989).Most pu blished studies of social influence resistance strate-gies have found modest but significant reductions, in compari-son with controls, in the onset and prevalence of cigarette sm ok-ing after training (see Botvin, 1986; Bukoski, 1986; Flay, 1985;Moskowitz, 1989; D. M. Murray, Davis-Hearn, Goldman, Pirie,& Luepker, 1988; Tobler, 1986, for reviews). A few studies havereported beneficial effects of the strategy in preventing or de-laying the onset of alcohol or m arij ua na use (Botvin , 1986; El-lickson & Bell, 1990; Hansen et al., 1988; McAlister, Perry, Kil-len, Slink ard, & M accoby, 1980; Pentz, Dwyer, et al., 1989).Student- or peer-led social influence resistance training in-terventions have achieved greater reductions in dr ug use, com-pared w ith interventions led by teachers (Botvin, Baker, Filaz-zola, & Bo tvin, 1990; K lep p et al., 1986; McAlister, 1983; D. M.Murray, Johnson, Luepker, & Mittelmark, 1984). This differ-ence may reflect greater fidelity in im pleme ntation of the cur-riculum by peer leaders (resulting in greater skill acquisition bystudents; Botvin et al., 1990), or the finding may reflect peerleaders' stimulation of classroom norms antithetical todrug use.A number of research issues remain to be addressed regard-ing the effects of social influence focused intervention. Onequestion is whether smoking prevention programs, withoutcontent specific to other drugs such as alcohol or marijuana,have effects on the use of these drugs. Some studies suggestthere may be a generalized effect of these prevention program son alcohol and marijuana use by subjects (G. M. Johnson et al.,1984; McAlister et al., 1980). A related question is raised byEllickson and Bell (1990), who sought to extend the social influ-ence model of smoking prevention to alcohol and marijuana.Results were mixe d. Modest reductions in drinking for studentsat three risk levels were observed immediately after the teen-ledversion of the program but disappeared at 1-year follow-up.Exposure to the curr icul um was associated with reductions insmoking amo ng baseline experimenters but increases in smok-ing among baseline smokers. Curriculum exposure was alsoassociated with reductions in both initiation and current use ofmarijuan a. T he investigators speculate that the apparent effec-tiveness of social influence approaches for tobacco and mari-


28/43

90 D. HAWKINS, R. CATALANO, AND J. MILLERjuana may reflect the generalized social norms against thosetwo substances but for alcohol social influence training is lesseffective because society has not developed a consensus againstits use.Biglan, Glasgow, et al. (1987), on the other hand, found nogeneralization of effects to alcohol or mar ijuan a use of a smok-ing refusal skills training program. Others have found that cog-nitive and interpersonal skills training interventions reducedtobacco use but had no effects on alcohol, marijuana, or otherdrug use (Gersick, Grady, & Snow, 1988). More research isneeded to determine whether preventing the onset of an earlybehavior in a sequence, such as smoking in the progression ofdrug use initiation, has effects on later behaviors in the se-quence.Another question is whether classroom-based social influ-ence resistance interventions have significant effects on adoles-cents at greatest risk for drug abuse. In most social influenceresistance studies, risk groups have been denned by differentlevels of baseline use, usually in smo king b ehavior (i.e., regulartobacco u sers, occasional tobacco users, and nonusers; Ellick -son, Bell, Thomas, Robyn, & Zellman, 1988). Although thisapproach can reduce smoking among students with parentsand friends who smoke (Botvin & Wills, 1985), few assessmentsare available of effects on those at greatest risk for drug abuse byvirtue of exposure to multiple risk factors earlier in develop-ment.Some social influence focused studies have looked at theeffects of preventive interventions on groups with special demo-graphic characteristics that may be related to high er risk. Bot-vin et al. (1989) addressed a special population of urban Afri-can-American youngsters with a smoking prevention programthat was based on life skills train ing usin g cognitive behavioraltechniques. Of several smoking outcomes exam ined, the onlysignificant effect observed was a smaller proportion of smokersat posttest in the treatment than in the control group on thebasis of adjusted means for smoking status in the past month.Some intervention effects were also observed for cognitive andattitude variables such as knowledgeof smoking consequencesand normative expectations.In another study designed to examine effects on a specificpopulation , Schink e, Botvin, et al. (1988) tested a social compe-tence/skills building intervention designed with cultural rele-vance for Native Am eric an adolescents. They found at posttestand 6-month follow-up that subjects who received the interven-tion improved more th an control subjects on measures of sub-stance use knowledge, attitudes and interactive skills, and self-reported rates of tobacco, alcohol and dru g use.The same group of investigators examine d still another spe-cial popula tion, children of blue-collar families, using a school-based social skills smoking prevention program (Schinke, Be-bel, Orlan di, & B otvin , 1988). Lower use rates were observedand validated among pupils who received the skills-based inter-vention (as compared with discussion-based groups and controlgroups) at 6, 12, 18, and 24 months follow-up. Although theyhave isola

risk and protective factors for alcohol and other drug problems in adolescence and early adulthood:...

Documents