Addictive Behaviors 31 (2006) 2038 – 2062
The role of drug use outcome expectancies in substance abuse risk:
An interactional–transformational model
Adam M. Leventhal a,b,*, Joy M. Schmitz b
b
a
Department of Psychology, University of Houston, Houston, TX 77204-5022, USA
Department of Psychiatry and Behavioral Sciences, The University of Texas-Houston, Houston, TX, USA
Abstract
While a variety of risk factors for substance abuse have been identified, the psychological mechanisms
underlying the transmission of risk is unclear based on studies using traditional risk-outcome research designs. The
present paper identifies drug use outcome expectancies as a common etiological mechanism involved in substance
abuse risk. Existing literature findings are reviewed and integrated according to an interactional–transformational
(IT) model of developmental psychopathology. This model identifies the preliminary (mediating) and secondary
(moderating) role of drug expectancies as important operations involved in the development of substance use
patterns. Advantages of the IT model over traditional trait-based or environmental models are discussed, along
with implications for intervention and future research.
D 2006 Elsevier Ltd. All rights reserved.
Keywords: Expectancies; Developmental psychopathology; Risk factor; Substance abuse
1. Introduction
A common approach in developmental psychopathology research involves examining the impact of
risk factors and protective influences on problematic behavior. In substance abuse, a variety of risk
factors have been identified, including intrapersonal characteristics such as low levels of education
* Corresponding author. Department of Psychology, University of Houston, Houston, TX 77204-5022, USA. Tel.: +1 713 563
4533; fax: +1 713 794 4730.
E-mail address: [email protected] (A.M. Leventhal).
0306-4603/$ - see front matter D 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.addbeh.2006.02.004
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2039
(Johnston & O’Malley, 1986), non-religiosity (Adlaf & Smart, 1985), impulsivity (Hartzler & Fromme,
2003; McCarthy, Kroll, & Smith, 2001; Moeller & Dougherty, 2002), emotional distress (Felix-Ortiz &
Newcomb, 1999), drug accepting attitudes and cognitions (Cloninger, Sigvardsson, & Bohman, 1988),
physiological and genetic predispositions (Schuckit, 1994), and the presence of comorbid psychopathology (Kilpatrick et al., 2000). In addition, numerous interpersonal risk factors have been identified,
including presence of substance-using family members and peers (Sussman, Dent, & Stacy, 1999),
exposure to popular media (Casswell, Gilmore, Silva, & Brasch, 1988), and various forms of physical
and sexual abuse (Cohen, Brown, & Smailes, 2001). Thus, a wide range of influences across numerous
domains may contribute to the development of substance abuse.
In general, risk factors for substance abuse have been studied independently, with little attention given
to identification of a common underlying vulnerability factor that may connect them. Having an
integrated model in which risk factors are linked together by a common mechanism would help to: (a)
explain how some seemingly unrelated risk factors produce similar substance use outcomes; (b) promote
a parsimonious explanation of substance abuse risk applicable to a wide range of circumstances; (c)
organize the already diverse vulnerability literature; (d) guide future research; and (e) provide a practical
intervention target.
This paper focuses on drug use outcome expectancies, which are beliefs regarding the anticipated
consequences of substance use, as a common vulnerability factor for substance abuse. We use the
interactional–transformational (IT) model of developmental psychopathology to integrate research
identifying the etiological role of drug expectancies in the development of substance abuse (Lewis,
2000). Because the IT model stresses the importance of risk factors that mediate and moderate the effects
of other vulnerabilities on maladaptive behavior, it provides a suitable framework by which expectancies
can be identified as significant etiological influences in the development and progression of substance
abuse. The current paper is organized into four sections. The first section presents general findings
uncovered by the risk-outcome model, the research paradigm in which vulnerability is typically studied.
In addition, approaches to developmental psychopathology are evaluated and their applications to
behavioral problems are considered. In the second section, the drug expectancy construct and the
developmental process of expectancy formation are reviewed. This section examines how expectancies
are thought to affect behavior and the methods that influence the mental construction of drug
expectancies. The third section presents findings that support an IT model of substance abuse
development in which drug expectancies mediate or moderate the impact of other risk factors. While
most of this literature examines developmental pathways in alcohol abuse, similar processes presumably
underlie vulnerability to other substances. The paper concludes by proposing a specific model upon
which IT concepts might be evaluated in clinical practice and research. Despite limitations, this new
approach offers significant advantages over existing risk-outcome paradigms.
2. Risk factors and their impact on the development of substance abuse
2.1. Risk factors identified using the risk-outcome approach
2.1.1. Family and social variables
The prediction of substance abuse based on family history has been a robust finding (Compton,
Cottler, Ridenour, Ben-Abdallah, & Spitznagel, 2002). Familial substance abuse can affect vulnerability
2040
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
in multiple ways. Physiological systems that modulate the pleasurable and aversive effects of drug
exposure may be transmitted genetically (McCarthy, Wall, Brown, & Carr, 2000; Uhl, Blum, Noble, &
Smith, 1993). Drug using family members may have positive attitudes regarding use, conferring
additional risk to offspring (Brown, Tate, Vik, Haas, & Aarons, 1999). Peer drug use may also increase
risk of substance abuse through social modeling influences (Zogg, Ma, Dent, & Stacy, 2004) and by
increasing the opportunity to obtain and consume drugs (Sussman et al., 1999; Voelkl & Frone, 2000).
2.1.2. Personality traits and behavioral tendencies
Substance abuse vulnerability can also arise from a variety of intrapersonal factors, such as
personality traits or behavioral tendencies. For example, drug use is associated with academic problems,
such as delinquency, school failure, and negative attitudes about school, suggesting that struggling
adolescents may engage in drug-seeking behavior as a means of escape (Voelkl & Frone, 2000).
Personality traits leaving individuals vulnerable to substance use have been described under the
constructs of impulsivity (Moeller & Dougherty, 2002), disinhibition (Patterson & Newman, 1993),
sensation-seeking (Kalichman, Tannenbaum, & Nachimson, 1998), type II (Cloninger, 1987), and lack
of future time perspective (Zimbardo & Boyd, 1999). Individuals with these traits typically show a
tendency to focus on rewards, learn less from punishment, and continue active responding in search of
reward even when passive avoidance of punishment is adaptive. This active, reward-seeking response
style combined with failure to learn from punishment might maintain drug-taking behavior in spite of
negative consequences. Some hypothesize that these traits are supported by and exert their influence on
behavior through a specific physiological motivational system (Gray, 1991).
2.1.3. Social and emotional functioning
Social and environmental factors in the form of maltreatment and victimization may lead to substance
abuse by modulating intrapersonal characteristics that are associated with vulnerability (Cohen et al.,
2001; Sussman et al., 1999). For example, maltreatment may lead to elevation of stress, depression, and
anxiety, which are proximal antecedents of drug use (Cooper, Russell, Skinner, Frone, & Mudar, 1992;
Kidorf & Lang, 1999). Victimization and feelings of insecurity may increase substance use as a way of
coping (Sussman et al., 1999), especially in individuals who lack an adequate and mature set of
psychological coping skills. Adolescents with inadequate social support and social skills may be at risk
for substance abuse (Spooner, 1999; Van Hasselt, Null, Kempton, & Bukstein, 1993). Spooner (1999)
hypothesizes that poor social functioning escalates risk because individuals who lack social skills tend to
be rejected by prosocial peers and associate with antisocial drug-using peers.
2.1.4. Distal factors
While emotional problems may acutely impact substance use patterns, distal factors have more of a
delayed and chronic impact on drug-taking behavior. A recent cross-national study demonstrated that
being male, non-married, of a low socioeconomic status (SES), and living in an urban setting were
associated with drug dependence (Furr-Holden & Anthony, 2003). This may be because these
individuals are less likely to stay in high school (Crum, Ensminger, Ro, & McCord, 1998) or attend
college (Johnston, O’Malley, & Bachman, 1987) and more likely to affiliate with delinquent peers
(Moss, Lynch, & Hardie, 2003). Other sociological variables, such as cultural norms (Moore, 1994) and
occupational status have been identified as possible risk factors for future substance use (Frank,
Jacobson, & Tuer, 1990). These distal factors have been argued to effect daily substance use patterns
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2041
indirectly or conditionally through other characteristics (Johnstone, 1994). More empirical research is
needed to isolate mechanisms of distal risk transmission.
2.1.5. Current substance use
Current drug use is a vulnerability factor for future substance abuse. Alcohol and cigarette use are
known to be significant risk factors for the initiation of illicit drug use (Bailey, 1992; Yu & Williford,
1992). More recently, use of alcohol, hallucinogens, or cocaine has been associated with current
stimulant use (Sussman et al., 1999). These findings support the notion that alcohol and cigarettes serve
as a risk factor for alcoholism and a bgatewayQ to illicit drugs. This phenomenon has been explained as a
manifestation of a problematic externalizing behavior in which all maladaptive behaviors, including
substance use, increase over time. However, this account is primarily descriptive, leaving the
psychological mechanisms underlying this process is unclear.
Overall, a variety of risk factors increases the chances of developing substance use problems. Riskoutcome investigations identify what, but not how, risk impacts drug-taking behavior. In order to better
understand the mechanisms underlying the risk-outcome relationship, models of developmental
psychopathology have been used to identify more specific causal pathways through which risk impacts
drug use (Lewis, 2000).
2.2. An interactional–transformational model of developmental psychopathology
Developmental psychopathology is defined as the study of maladaptive behaviors and processes
across time (Lewis, 2000). Lewis (2000) proposes three broad models of developmental psychopathology. The trait model hypothesizes that stable individual characteristics result in the expression of certain
behaviors while environment has little impact on behavior. Traits are seen as stable tendencies that
continually impact psychopathological risk throughout development by increasing vulnerability or
serving as protective factors. For example, neuroticism is considered to be an important risk factor for
the development of anxiety disorders (Chorpita & Barlow, 1998). While this approach is appealing
because of its simplicity, it does not account for changes in behavior across time and situations. In
contrast, the environmental model proposes that behavior results from environmental contingencies and
individuals are relatively passive in these transactions. Environmental models account for data showing
that maltreated children develop more psychopathology than non-maltreated individuals (Cohen et al.,
2001), however they do not explain instances in which environmental contingencies fail to change
behavior as predicted. For example, it has been demonstrated that certain resilient individuals develop
little or no psychopathology despite experiences of significant adversity (Luthar, Cicchetti, & Becker,
2000). Despite shortcomings of both, trait-based and environmental models have been attractive to
researchers because they fit the traditional risk-outcome paradigm. In contrast, the interactional–
transformational (IT) model takes into account both the inconsistency and stability of behavior across
and within contexts and lends itself to a different research approach.
The IT model proposes that psychopathology is caused by an interaction of multiple influences (Ladd
& Burgess, 2001). Interactive processes are believed to underlie diathesis–stress models of emotional
disorders, in which the presence of vulnerability factors, such as depressive attributional-style, is
believed to produce the onset of affective distress only in the presence of stressful life events (Conley,
Haines, Hilt, & Metalsky, 2001). The notion of transformation refers to changes in environmental or
intrapersonal risk factors as a result of previous conditions. Hence, one risk factor can transform another,
2042
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
which in turn modifies one’s vulnerability to developing psychopathology. For example, recent evidence
indicates that childhood maltreatment damages emotion regulation abilities, causing increased risk of
anxiety and depression (Maughan & Cicchetti, 2002). IT models build on traditional risk-outcome
approaches by providing causal explanation of the mechanisms through which (mediation) and the
conditions under which (moderation) the risk-outcome relationship holds. The IT approach of
developmental psychopathology lends itself to two specific research paradigms: (1) moderational
models, in which the relation between risks and behavior problems is amplified or reduced by other
factors; and (2) mediational models, in which risk factors transform other vulnerabilities that cause
psychopathology.
Recent efforts in developmental psychopathology research have employed IT models to better
understand the mechanisms underlying the development of behavioral problems (see Margolin and
Gordis, 2000 for a review of mediators and moderators of the behavioral consequences of interpersonal
violence). A growing literature indicates that drug expectancies are important vulnerability factors that
both mediate and moderate the effect of other risks on the development of substance abuse. Before
evidence for specific expectancy-based IT pathways is reviewed, the drug expectancy construct is
examined.
3. Drug expectancies and their association with substance abuse
3.1. The expectancy construct
Based on social learning perspectives, outcome expectancy theory proposes that people engage in
certain behaviors based on their beliefs of the behavior’s specific reinforcing effects (Bandura, 1977). In
applying expectancy theory to substance use, it is assumed that drug-taking behavior is motivated by the
desire to attain particular outcomes associated with drug consumption. Typically, drug expectancies have
been dichotomized into positive and negative forms. Positive expectancies (e.g., bI expect to feel a sense
of excitement and pleasure if I use cocaineQ) are thought to motivate increases in drug-taking behavior,
whereas negative expectancies (e.g., bI expect to feel lousy the day after drinkingQ) represent motivation
to restrict substance use (Oei & Morawska, 2004). The major component of expectancy theories of
substance abuse is that variation in the valence and strength of expectancies explains individual
differences in substance use patterns and intra-person variation in drug use across contexts. Informationprocessing approaches view drug expectancies as informational structures in long-term memory
(Goldman & Darkes, 2004). Within this framework, the content of these expectancies is thought to be
less important than their accessibility (i.e., likelihood to be activated from memory) in drug-using
situations (McCusker, 2001; Rather & Goldman, 1994).
Although there is some research on the content of marijuana and stimulant expectancies (e.g., Aarons,
Brown, Stice, & Coe, 2001), most of the research to date has focused on the specific components of
alcohol expectancies (see Jones, Corbin, & Fromme, 2001 for a review). Common positive expectations
of drinking outcomes include social assertion, enhanced cognitive and motor functioning, sexual arousal,
reduction of negative affect and distress, and physical and social pleasure (Jones et al., 2001). Negative
alcohol expectancies involve perceptions that drinking causes sickness, sadness, dizziness, sleepiness,
and dangerous behavior (Rather & Goldman, 1994). The most prominently used measure of alcohol
expectancies is the Alcohol Expectancy Questionnaire (AEQ: Brown, Christiansen, & Goldman, 1987).
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2043
The AEQ focuses on positive expectancies and has six subscales related to enhancement of global
positive outcomes, social and physical pleasure, sexual experience, power and aggression, social
assertiveness, and relaxation. Early research by Brown and colleagues (Brown, 1985a,b; Brown,
Goldman, Inn, & Anderson, 1980) indicated that specific expectancy scales were related to different
substance use patterns. They found that light drinking was associated with global positive expectancies,
while heavy drinking patterns were related to expectations of tension reduction, aggression, enhanced
sex and social pleasure. Abusive drinking patterns were more likely to be related to global positive
expectancies and social assertion expectancies. More recent investigations have failed to support
psychometrically distinct scales of the AEQ, suggesting that it is most useful to think of this instrument
as measuring a general belief about alcohols desirable effects (Leigh, 1987, 1989a,b). Items measuring
expectations of negative outcomes also appear to converge upon a common construct (Leigh & Stacy,
1993). While some research suggests that the expectancy subscales have demonstrated poor discriminant
validity (Leigh, 1987, 1989a,b), unidimensional expectancy measures robustly predict substance use
patterns, although positive expectancies are typically a more powerful motivator of substance use than
negative expectancies (Jones et al., 2001; Leigh & Stacy, 1993).
3.2. Developmental trajectory of expectancy formation
Research from several studies indicates that drug expectancies change along with developmental
maturation (Dunn & Goldman, 1998; Smith, Goldman, Greenbaum, & Christiansen, 1995). Exposure
to substance-using models in peers, family members, and the media is likely to facilitate
construction of drug expectancies even before use is first initiated (Brown, Creamer, & Stetson,
1987; Casswell et al., 1988). This information tends to become stronger, more developed in longterm memory, and more positive in valence as one initiates and maintains use (Dunn & Goldman,
1998; Sher, Wood, Wood, & Raskin, 1996; Smith et al., 1995). Several studies have shown that as
adolescents become older and begin drinking, the positive effects of alcohol become more salient
(Dunn & Goldman, 1998; Zogg et al., 2004). In contrast, thoughts about cocaine tend to become
more negative with increased age (Bridges et al., 2003). However, research of the developmental
trajectories of expected outcomes of stimulant, hallucinogen, and opioid use requires further
clarification. On the whole, important changes in the cognitive organization of drug expectancies occur
throughout development.
To summarize, drug expectancies are cognitive representations that can be conceptualized in terms of
their effects on behavior or as distinct memory structures that influence decisions to consume drugs
through memory activation. Expectancies are formed through direct experience with substance use or
through social processes. Furthermore, drug expectancies change throughout development, which
indicates that environmental and intrapersonal experiences can transform expectancies. These findings
raise further questions: How does this process impact the development of substance use problems and
how do expectancies enhance the impact of other risk factors on drug use outcomes?
4. The role of expectancies in substance abuse risk
IT models of developmental psychopathology point to two possible research designs. The first
involves examining whether expectancy variables directly or indirectly mediate the relation between
2044
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
risk and drug use. This method tests the hypothesis that a vulnerability factor transforms expectancies,
which in turn impacts drug-taking behavior, indicating the mechanisms through which risk factors
influence behavior (see Fig. 1, point A). Partial mediation may also be present, in which a risk factor
modifies drug-taking behavior both directly and through expectancies (see Fig. 1, point B). The
second paradigm explores whether expectancy variables moderate the effect of risk on drug use. This
model tests the hypothesis that a risk factor affects drug-taking behavior with greater or lesser effect
depending on one’s expectancies, demonstrating mechanisms under which risk factors influence
behavior (see Fig. 1, point C). As shown in Table 1, findings in the existing literature document the
mediational and moderational function of expectancies across numerous domains of risk. We
summarize several pertinent examples of IT pathways demonstrated in the literature to illustrate how
expectancies operate in substance abuse risk.
4.1. Family and social variables
The interaction of positive expectancies and familial history of alcoholism have been shown to predict
problem drinking symptoms (Conway, Swendsen, & Merikangas, 2003; Ohannessian, Hesselbrock,
Tennen, & Affleck, 1994), suggesting several pathways by which family history influences drug use.
Individuals with substance abusing parents may see a greater opportunity to take drugs, which would
promote drug-seeking behavior in those who believe that substance use produces desired outcomes.
A. Full Mediational Pathway
Risk
Factor
Maladaptive
Expectancies
Substance
Use
B. Partial Mediational Pathway
Maladaptive
Expectancies
Risk
Factor
Substance
Use
C. Moderational Pathway
Maladaptive
Expectancies
(amplifier)
Risk
Factor
Substance
Use
Fig. 1. Mediational and moderational pathways in expectancy-based interactional–transformational models of substance use
development.
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2045
These children may also have family members with accepting attitudes about substance use, making
them more likely to act on their expectancies.
In addition to exerting a moderating effect, drug expectancies may also mediate the relationship
between family and social factors and substance use. Family and peer attitudes about drinking (Ouellette,
Gerrard, Gibbons, & Reis-Bergan, 1999; Wood, Read, Palfai, & Stevenson, 2001), the perception that
one’s friends use drugs (Hine, McKenzie-Richer, Lewko, Tilleczek, & Perreault, 2002; Wood et al.,
2001), perceived pressure from peers to use substances (Wood et al., 2001), and normative beliefs about
substance use (Fearnow-Kenny, Wyrick, Hansen, Dyreg, & Beau, 2001) may alter drug expectancies and
subsequent drug use patterns. Family and peers may promote religiosity, which has been shown to
reduce substance use though promoting negative expectancies (Galen & Rogers, 2004). Taken together,
findings from these studies suggest that social influences, such as familial substance abuse, may enhance
the development of drug expectancies, which in turn impact drug-taking behavior.
In combination with many social modeling influences, heritability is a major component of family
history risk (Merikangas, 2002). To the extent that offspring of substance abusers inherit nervous
systems that enhance the pleasurable effects and diminish the aversive consequences of drug
intoxication, as some research suggests (Uhl et al., 1993), this, in turn, may promote positive
expectancies and motivate substance use. Alternatively, individuals can inherit biological characteristics
that lower the risk for drug use (Wall, Thomasson, Schuckit, & Ehlers, 1992). McCarthy et al. (2000)
found that women who experience facial flushing, tachycardia, hypotension, nausea, and vomiting in
response to moderate alcohol drinking were less likely to expect tension reduction, enhanced sexuality,
and general positive outcomes from drinking, which further predicted lower levels of alcohol
consumption. They also demonstrated that these women were less likely to act on positive alcohol
expectancies than women with typical responses to alcohol. On the whole, results from several
investigations indicate that expectancies interact with or mediate the impact of risk due to biological and
social components of a family history of substance abuse.
4.2. Personality traits and behavioral tendencies
Models identifying the importance of impulsive personality traits in the development and maintenance
of drug problems (Cloninger, 1987; Moeller & Dougherty, 2002; Patterson & Newman, 1993) have
recently been extended to include the mediational and moderational role of expectancies (McCarthy et
al., 2000; Sher, Walitzer, Wood, & Brent, 1991). For example, the acquired preparedness model proposes
that impulsivity promotes the development of positive expectancies and inhibits generation of negative
expectancies, resulting in a greater likelihood of substance abuse. Several studies have supported this
mediational model in predicting marijuana use (Vangsness, Bry, & LaBouvie, 2005), cocaine use (Stacy,
Newcomb, & Bentler, 1995), and alcohol use in college student (McCarthy et al., 2000) and adolescent
populations (Barnow et al., 2004). In a similar vein, Powell, Bradley, & Gray (1992) showed that the
association between impulsivity and cue-elicited craving in treatment-seeking opiate abusers was
mediated by positive opiate expectancies. Kalichman et al. (1998) examined the influence of sensation
seeking traits on pre-sex substance use in gay and bisexual men and showed that this relation was
partially mediated by expectations that drug and alcohol use enhances sexual outcomes. There is also
evidence for a moderational pathway involving impulsivity and expectancies in which disinhibited
individuals are at greater risk of substance abuse if they hold more positive expectancies (McCarthy et
al., 2000).
2046
Family and social variables (sample)
Ouellette et al. (1999) (adolescents)
Conway et al. (2003) (adults)
Simons-Morton (2004) (sixth grade students)
Substance(s)
Risk/protective factor
Expectancy variable
Mediation?
Moderation?
Alcohol
Alcohol
Alcohol
Parental and peer drinking
Family history of alcoholism
Parental attitudes about substance
use and disruptive behavior
Genetically transmitted sensitivity to alcohol
General positive expectancies
Global positive changes
General expectancies
+
+
NR
NR
NR
+
General positive expectancies
Enhanced sexuality
Tension reduction
Enhanced courage
General negative expectancies
Cognitive–behavioral impairment
Increased risk and aggression
Negative self-perception
Enhanced performance
Tension reduction
General expectancies
Negative affect control
Social costs
Weight control
General expectancies
Negative affect control
Social costs
Weight control
+
+
+
+
+
+
NR
NR
+
McCarthy et al. (2000) (Asian American
female college students)
Alcohol
Voelkl and Frone (2000) (high school students)
Alcohol and
marijuana
Tobacco
Hine et al. (2002) (adolescents)
Ease of use
Peer tobacco use
Parent tobacco use
Personality traits and behavior tendencies (sample)
McCarthy et al. (2001) (male undergraduates)
Katz et al. (2000) (college students)
Alcohol
Illicit drugs
Disinhibition
Sensation seeking
Low social conformity
Alcohol
Sensation seeking
Low social conformity
Trudeau et al. (2003) (young adolescents)
Alcohol, tobacco,
and marijuana
Assertiveness
Decision making skills
General
General
General
General
General
General
General
General
General
General
General
positive expectancies
beneficial outcomes
risky outcomes
beneficial outcomes
risky outcomesa
beneficial outcomes
risky outcomesa
beneficial outcomes
risky outcomesa
negative expectanciesa
negative expectanciesa
+
NR
NR
+
+
+
NR
NR
NR
NR
NR
NR
NR
NR
+
NR
NR
NR
NR
NR
NR
NR
NR
NR
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
Table 1
Studies examining expectancy-based IT pathways to substance use
Copeland and Carney (2003) (undergraduate
women)
Kalichman et al. (1998) (gay and bisexual
male adults)
Stacy et al. (1995) (community volunteers)
Social/emotional functioning (sample)
Griffin et al. (2001) (adolescents)
Mooney and Corcoran (1989) (rural male
undergraduates)
Stacy et al. (1995) (community volunteers)
Poor dietary restraint
Appetite and weight control
+
NR
All substances
Sensation seeking
Enhanced sexual outcomes
+
NR
Cocaine
Boredom susceptibility
General positive expectancies
+
NR
Alcohol and
tobacco
Alcohol
Social competence
Social benefit expectancies
+
NR
Low assertiveness
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
+
+
NR
Cocaine
Depression
Global positive outcomes
Enhanced sexuality
Social and physical pleasure
Enhanced social assertiveness
Relaxation and tension reduction
Increased arousal and aggression
Global positive outcomes
Enhanced sexuality
Social and physical pleasure
Enhanced social assertiveness
Relaxation and tension reduction
Increased arousal and aggression
General positive expectancies
Alcohol
Low assertiveness
+
+
+
+
+
+
+
Other risk factors (sample)
Tapert et al. (2003) (substance use disordered
adolescents)
Levy and Earleywine (2003) (college students)
McCarthy et al. (2002) (treated young adults)
Alcohol
Language skills
Global positive changes
+
NR
Alcohol
Alcohol
Poor studying attitudes
Educational attainment
General positive expectancies
General positive expectancies
+
+
NR
NR
Current use (sample)
Hine et al. (2002) (adolescents)
Tobacco
Current tobacco use
General expectancies
Negative affect control
Social costs
Weight control
General positive expectancies
Stacy et al. (1995) (community volunteers)
Cocaine
Polydrug use
+
+
+
NR
NR
NR
NR
NR
+ Denotes evidence of an effect. Denotes evidence of no effect. NR denotes that effect was not reported. If not otherwise stated, IT pathways involve the transmission of risk in which positive
relationships between risk factors and substance use are amplified or positively mediated by expectancy variables.
The current table provides a general overview of mediational and moderational effects of expectancies on substance use risk. Only relevant findings are included. For more detailed
understanding, reader should consult original papers.
a
Denotes IT pathways involving protective factors in which positive relationships between protective factors and decreased substance use are amplified or positively mediated by expectancy
variables.
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
Mooney and Corcoran (1989) (rural
female undergraduates)
Tobacco
2047
2048
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
4.3. Social and emotional functioning
Cooper and colleagues found that individuals experiencing stressors or negative affect were more
likely to drink if they reported positive alcohol expectancies about mood regulation (Cooper, Frone,
Russell, & Mudar, 1995; Cooper et al., 1992). Other investigators examining the relation between
particular types of expectancies and specific areas of social and emotional functioning report that men
with social anxiety drink more during a socially stressful period if they expect alcohol to increase their
assertiveness (Kidorf & Lang, 1999). For example, Mooney and Corcora (1989) found that for less
assertive college students, the expectation that alcohol would facilitate social assertion predicted
drinking behavior; however, for high assertive students, drinking patterns were not predicted by the
expectation of enhanced social assertion. Trudeau et al. (2003) found that assertive students were more
likely to develop negative cigarette, alcohol, and marijuana expectancies, which prevented substance
use initiation. In addition, they reported that students with good decision-making skills tended to
develop stronger negative expectancies, and were less likely to start using substances. Thus, children
with these skills may be better able to see the negative long-term effects associated with substance use
than those lacking decision-making capabilities. Taken together, these findings indicate that specific
social and emotional vulnerabilities may interact with particular expectancies to influence drug-taking
behavior.
4.4. Other distal factors
Evidence suggests that individuals with above average language abilities are more prone to alcohol
involvement and alcohol dependence symptoms if they hold positive alcohol expectancies (Tapert,
McCarthy, Aarons, Schweinsburg, & Brown, 2003). Theoretically, these individuals have the capacity to
engage in higher-level cognitive processes, which may enhance their ability to foresee future positive
outcomes of alcohol use. McCarthy and colleagues demonstrated that the effect of educational
achievement on drinking behavior was partially mediated by the expectations that drinking causes global
positive outcomes (McCarthy, Aarons, & Brown, 2002). However, more specific studies are necessary to
replicate these findings and investigate the theoretical mechanisms behind such pathways.
4.5. Current use
A large base of findings indicates that present substance use causes future use through the
enhancement of positive expectancies (Hine et al., 2002; Sher et al., 1996; Smith et al., 1995; Willner,
2001). For example, alcohol use may serve as a bgatewayQ to illicit drug use through the formation of
positive drug expectancies, although this hypothesis has not been tested longitudinally (Willner, 2001).
Longitudinal studies have used structural equation modeling techniques to demonstrate reciprocal
relationships between expectancy formation and alcohol use in novice drinkers (Sher et al., 1996; Smith
et al., 1995; Willner, 2001). In adolescents and undergraduates, the reciprocal relationship appears to be
a positive feedback loop between positive expectancies and use (Sher et al., 1996; Smith et al., 1995).
However, a 9-year prospective study of individuals who aged from 18 years at time one to 27 years at
time two found no evidence of a prospective effect of drug use on expectancies (Stacy, Newcomb, &
Bentler, 1991). Divergence in outcomes across different aged samples may indicate that the effect of use
on expectancy formation varies as a function of developmental stage (Sher et al., 1996).
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2049
The expectancy-based IT pathways reviewed in this paper demonstrate that a variety of risk factors
operate in a similar fashion, either by being mediated or moderated by the presence of drug use outcome
expectancies. In some cases, the effect of a single vulnerability factor on substance use can be both
mediated and moderated by expectancies (McCarthy, Brown, Carr, & Wall, 2001; McCarthy et al.,
2001). This suggests that the mediational and moderational developmental processes may be linked,
combining to play a major role in substance abuse risk. The following section explores the joint
influence of mediational and moderational processes.
5. An expectancy-based IT model of substance abuse risk
5.1. A two-process framework
Cognitive-developmental theories of anxiety and depression propose that the structural relation
between risk, cognitive, and clinical variables differs as a function of developmental stage (Chorpita &
Barlow, 1998; Cole & Turner, 1993). In early childhood, risk factors may lead to the construction of
maladaptive cognition, which mediates the impact of risk on emotional expression. The moderational
effects of cognition on risk-outcome relations are believed to operate during later stages of development
(e.g., the effect of risk on emotional functioning is amplified by the presence of cognitive vulnerability).
That is, the presence of early risk may promote a maladaptive cognitive framework (i.e., mediational
pathway). Later in adolescence and adulthood, cognitive vulnerability may begin to operate as an
amplifier for risk factors (i.e., moderational pathway).
The normative path of cognitive–psychological development supports this dual process approach to
developmental psychopathology. In typical development, symbol formation develops at the end of
infancy (Piaget, 1962; Vygotsky, 1978), the emergence of logical or scientific thought occurs during
middle childhood (Inhelder & Piaget, 1969; Vygotsky, 1986), and the construction of a conception of
self is a slow process that develops from infancy through adolescence (Stipek, Recchia, & McClintic,
1992). These developmental milestones mark changes in the overall capacity for abstract thought and
suggest that children do not have the capability to construct stable abstract cognitions (like those that
would promote psychopathology) until later stages of development. Therefore, risk factors may not be
able to interact with cognitive factors to produce psychopathology at earlier ages because risk-enhancing
cognitions are not readily accessible. However, in later stages of development when risk-enhancing
cognitions become more solidified, they can become activated in the presence of other risk factors to
enhance vulnerability to psychopathology.
There is also empirical support for a two-process model. In two studies, Cole and Turner
comparatively evaluated moderational and mediational models of depression in a sample of children and
adolescents (Cole & Turner, 1993; Turner & Cole, 1994). They found support for a mediational model in
which a non-cognitive risk factor (i.e., unpleasant events) influenced a cognitive risk factor (i.e., negative
attributional style), which in turn predicted depression. Under the dual process framework, they also
hypothesized that a moderational model would operate only in the older adolescents and would not be
observable in the younger children. Their hypothesis was supported by evidence of an
Age Event Cognition interaction.
The notion that differential pathways exist across developmental levels can be applied to explain the
development of substance abuse. Although research has not typically examined whether the statistical
association (i.e., moderation vs. mediation) between expectancies and risk-outcome pathways differ as a
2050
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
function of developmental stage, this may be the case. We integrate these ideas into a two-process
conceptual model of expectancy-implicated pathways to substance abuse (see Fig. 2).
The model proposes a preliminary process in which risk factors foster the construction of expectancies
in early stages of development before drug use begins, a time sometimes referred to as the bcritical
periodQ (Miller, Smith, & Goldman, 1990). Believing that substance use can produce desired outcomes
serves as a motivator for initial substance use. As part of this preliminary mediational process, specific
substance use experiences provide first hand information leading to the development of stronger
expectancies that become stored in memory and further influence drug-taking behavior. Indeed,
substance use has been shown to prospectively predict the development of more positive expectancies in
younger (Sher et al., 1996; Smith et al., 1995) but not older individuals (Stacy et al., 1991). In a
secondary process, established expectancies play a moderational role by setting up cognitive conditions
that, in the presence of risk factors, increase substance use. The vulnerability factors originally involved
in mediating processes remain present, however their influence on substance use behavior may vary
depending on the strength and valence of existing expectancies (McCarthy et al., 2001; McCarthy et al.,
2000; Stacy et al., 1991). Moreover, the influence of new vulnerability factors on substance use, which
were not involved in the preliminary process, are also enhanced under conditions of frequent activation
of strong positive expectancies and infrequent activation of weak negative expectancies.
To illustrate the utility of this model, consider the IT processes that may take place when family
history of alcoholism acts as a risk factor for alcohol abuse. An adolescent observes his alcoholic father
drinking at parties and after work and begins to form the expectation that alcohol produces pleasure and
relieves stress, the first stage of the preliminary (mediational) process. Later, he attends a party where he
sees friends enjoying themselves while drinking. Because of his positive expectations, he too decides to
drink, starts enjoying himself, and feels less anxious. Consequently, his expectancies are confirmed,
becoming stronger and more accessible in memory and his motivation to drink at parties and in other
social situations increases. A secondary (moderational) process may later begin to operate that involves
Preliminary Process
Secondary Process
maintenance of prior risk
risk
factors
onset of new vulnerabilities
memory storage
maladaptive
expectancies
enhancing
solidification
strengthening
maladaptive
expectancies
risk
factors
substance use
substance use
Fig. 2. A general expectancy-based model of substance use development proposing that the development of substance use
problems is driven by two processes. In a preliminary process, the influence of risk factors on drug taking is mediated by
maladaptive expectancies. Substance use may help confirm and solidify expectancies in memory. In a secondary moderational
process, the impact of risk factors on substance use is amplified by the presence of maladaptive expectancies.
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2051
his father’s permissive attitude about high school drinking, a component of the original risk factor of
family history of abuse. Under these conditions, this adolescent is likely to act on his expectancies
without fear of being punished. However, other adolescents who hold more negative alcohol
expectancies may be less likely to take advantage of their parents’ permissive attitudes about adolescent
drinking. The present example illustrates how a single risk factor (family history of substance abuse)
might operate in preliminary and secondary IT processes. However, in many instances, multiple risk
factors may operate individually or collectively, thus the two-process model accounts for a variety of
situations and specific vulnerabilities. The model also isolates expectancies as a common etiological
component in substance abuse vulnerability, and an appropriate target for intervention.
5.2. Implications for intervention
An expectancy-based IT model of substance abuse vulnerability points to four possible intervention
alternatives: preventing the onset of risk, transforming already present risk factors, altering expectancy
formation trajectories, and modifying current maladaptive expectancies. The first two alternatives are
typically the less efficient means of reducing drug use behavior, especially when risk factors are distal
and final (e.g., parental divorce, neighborhood climate, biological predispositions) (Trudeau et al., 2003).
However, interventions targeting expectancies have shown some promise (Cruz & Dunn, 2003; Darkes
& Goldman, 1993, 1998; Dunn, Lau, & Cruz, 2000; Musher-Eizenman & Kulick, 2003).
Most of the drug expectancy intervention literature has explored the use of brief alcohol expectancy
challenges (Cruz & Dunn, 2003; Darkes & Goldman, 1993, 1998; Dunn et al., 2000; Musher-Eizenman
& Kulick, 2003). These challenges typically begin by asking participants to consume alcohol or placeboalcohol beverages, with the drinkers being blind to their condition. The intent is to provide drinkers with
a direct experience of the true pharmacological effects of alcohol in contrast to the effects caused by
outcome expectancies. Through group discussion, drinkers learn that alcohol is a central nervous system
depressant with known pharmacological effects (e.g., sleepiness, nausea, and dizziness), whereas many
of the desired behavioral effects are due to expectations more than pharmacology (see Darkes and
Goldman, 1993 for a review of their protocol). Expectancy challenges are effective in weakening
positive alcohol outcome expectancies and strengthening negative alcohol outcome expectancies in
memory (Dunn et al., 2000). It is unknown whether challenges produce lasting effects; however, they
have been shown to reduce alcohol intake for up to 6 weeks beyond active treatment (Darkes &
Goldman, 1993, 1998; Dunn et al., 2000; Musher-Eizenman & Kulick, 2003). Moreover, decreases in
drinking parallel decreases in the strength of positive alcohol outcome expectancies in subjects
participating in expectancy challenges (Darkes & Goldman, 1993; Dunn et al., 2000).
Expectancy challenges for children aim to prevent the development of positive expectancies, thereby
interrupting the preliminary (mediational) process in substance abuse risk (Cruz & Dunn, 2003).
Childhood challenges rely solely on instructional strategies to teach pre-adolescents about the biological
effects of alcohol and dispel myths about drinking. Initial findings have demonstrated that this
intervention successfully modified 4th graders’ alcohol expectancies (Cruz & Dunn, 2003). Future
studies will examine whether these cognitive changes persist throughout adolescence and reduce highschool alcohol consumption.
Beyond alcohol, brief expectancy-based interventions have been shown to cause reductions in
drinking, cigarette smoking, and use of other substances, such as marijuana (D’Amico & Fromme, 2000;
Marlatt et al., 1998). While single-session prevention programs are efficient at reducing short-term drug-
2052
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
taking behavior, these positive changes may not be maintained over longer durations (D’Amico &
Fromme, 2000). On the other hand, long-term interventions have been quite successful in changing
expectancies and preventing substance use (Kivlahan, Marlatt, Fromme, & Coppel, 1990). For example,
the Alcohol Skills Training Program involves 8 weekly, 90-min sessions that emphasize relapse
prevention skills, include a placebo drinking session in a simulated tavern to identify alcohol outcome
expectancies and to challenge inappropriate attributions of pleasurable effects to drug rather than setting,
and present information from research on alcohol expectancies. In one study, this program demonstrated
a 38.5% reduction in weekly alcohol consumption in comparison to 21.6% for a lecture based alcohol
education program and 16% for assessment only (Kivlahan et al., 1990).
5.3. Limitations
Although a promising research area, there are several notable limitations of many expectancy-based
IT investigations and the general two-process model. For example, interactional and transformational
models of developmental psychopathology have been criticized for being too broad (Lewis, 2000).
Within this paradigm, there are literally thousands of combinations of risk factors, mediators, and
moderators that can influence the development of psychopathology. Some may argue that the addition of
mediators and moderators reduces the parsimony of simpler trait-based or environmental models.
Although IT explanations can be more complex than risk factor models, we argue that environmental or
trait approaches do not adequately account for the developmental processes that result in substance use
and abuse. Unlike IT-models, risk-outcome approaches fail to explain: (1) why people with similar risk
factors often do not have the same vulnerability to developing substance abuse; (2) how distal risk
factors can impact immediate drug use behaviors; and (3) why so many different classes of risk factors
can result in the same developmental outcome of substance abuse. Thus, the IT model’s drawbacks,
including complexity and breadth, are outweighed by its capability to adequately account for the various
paths to substance abuse. Overall, the IT model, with its focus on a common etiological mechanism,
offers an efficient and more focused approach than risk-outcome models.
A major limitation of many expectancy-based IT investigations is the over-reliance on self-report
assessment, which can cause a method bias that spuriously inflates effect sizes (Shadish, Cook, &
Campbell, 2002). Many have argued that expectancies impact drug-taking behavior both through
implicit cognition, which involves automatic processes that occur in the absence of conscious
deliberation, and explicit cognitive processes that require deliberate retrieval of information (McCusker,
2001; Roehrich & Goldman, 1995; Weingardt, Stacy, & Leigh, 1996; Wiers, Van Woerden, Smulders, &
De Jong, 2002). In fact, recent evidence indicates that both explicit and implicit cognition contribute
uniquely to the prediction of substance use behavior (Cox, Hogan, Kristian, & Race, 2002; Waters et al.,
2003; Wiers et al., 2002), however relatively few studies have examined implicit expectancies from a
developmental perspective. Because little is known about implicit processes and their relation to specific
risk factors, future IT studies of substance abuse risk use may wish to employ assessment techniques that
measure implicit cognition (e.g., Stacy & Newcomb, 1998).
5.4. Future research
Although there has been increased interest in the investigation of expectancy-based IT pathways,
some major gaps in the literature remain. These include examining pathways to bhard drugQ abuse (e.g.,
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2053
stimulants, opioids, and hallucinogens), identifying the role of developmental stages in substance abuse
risk, evaluating trajectories involving protective influences, and extending the IT model. Using the IT
framework can help clarify how research on substance abuse risk might proceed and be utilized.
All expectancy-based IT pathways contain three classes of variables: (i) a primary risk/protective
variable; (ii) an expectancy variable; and (iii) a drug use outcome variable. However, different
developmental trajectories might incorporate variables that tap different constructs within the three
classes. As discussed earlier, the primary risk variable could index personality constructs, social and
emotional functioning, other distal factors, or could even be a protective factor, such as assertiveness
(Trudeau et al., 2003). The expectancy variable could be represented as global positive or negative
expectancies or more specific cognitions, such as drug use to enhance sexual performance (Kalichman et
al., 1998). The outcome variable could be either initiation of use, duration and frequency of use, severity
of abuse or dependence, or comorbid substance use and mental disorder status.
Although the content of variables may vary, the functional relationship among variables in
expectancy-based IT pathways does not differ across pathways. That is, regardless of the substance of
interest, primary risk/protective factor, type of expectancy, or outcome variable, the IT model proposes
that expectancies function to mediate and moderate relationships between primary risk factors and
outcomes. Thus, the IT model provides a general framework for investigations of substance abuse risk.
5.4.1. bHard drugsQ and IT models
The bulk of expectancy-based IT model research has examined the impact of expectancies on alcohol,
marijuana, and tobacco use development (bnon-hard drugsQ). Very few studies have looked at drug
expectancies in opiate, stimulant, hallucinogen, and other bhard drugQ use. The distinction could be
critical for several reasons and necessitates further research of expectancy-based IT pathways in hard
drug abuse.
First, risk factors for hard drug abuse may be different from those associated with marijuana,
tobacco, and alcohol use disorders. Kilpatrick et al. (2000) demonstrated that for a specific risk factor,
the degree of substance abuse risk varies depending on the drug. For example, family history of any
drug use disorder showed approximately a 2-, 4-, and 8-fold increase in risk for developing alcohol,
marijuana, and hard drug use disorders, respectively. This study also demonstrated that certain
variables that were associated with increased risk for one substance were not risk factors for other
substances. For instance, male adolescents were at greater risk of marijuana and alcohol/dependence
than women; however, gender did not associate with hard drug abuse/dependence (Kilpatrick et al.,
2000).
Second, the expectancy construct differs between hard drugs and alcohol, marijuana, and tobacco.
Although there is modest overlap between hard and non-hard drug expectancies (e.g., both include
global positive and negative effects as well as relaxation and tension reduction expectations; Schafer &
Brown, 1991), some expected outcomes of hard drug use are unique to the particular substance. The
motivational checklist that assesses opiate expectancies contains several items unique to opiate abuse,
including relief of physical withdrawal symptoms, pleasure of a brush,Q and fear of criminal conviction
(Powell et al., 1992). The Cocaine Effect Expectancy Questionnaire contains items exclusive to cocaine
use, such as bI grind my teeth when I’m on cocaineQ (Schafer & Brown, 1991).
Third, the relation between outcome expectancies and type of substance use outcome may differ
depending on the substance. Schafer and Brown (1991) compared the association between use status and
cocaine and marijuana expectancies among college students. They found that a dimension of marijuana
2054
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
expectancies characterized by global negative effects discriminated primarily between non-users and all
use groups (i.e., infrequent, recreational, and regular users). However, a similar dimension of
expectancies for cocaine use demonstrated the ability to distinguish between infrequent and regular
cocaine users.
The distinction between hard and non-hard drugs has several implications for IT models. Primarily,
pathways modeling hard versus non-hard drug use outcomes might incorporate different constructs in
their representation of the three classes of variables in expectancy-based IT models. Based on previous
findings reviewed above (Schafer & Brown, 1991; Kilpatrick et al., 2000), models examining the
development of cocaine abuse might incorporate family history of drug dependence as a primary risk
factor, teeth grinding as an expectancy variable, and heaviness of use as an outcome variable. Pathways
modeling alcohol use outcomes would be less likely to include these variables and might focus more on
family history of alcohol use as a primary risk factor (Conway et al., 2003).
The functional relationship among variables in expectancy-based IT pathways does not differ across
substances. Indeed, several studies have identified the mediating role of expectancies in predicting
outcomes in cocaine (Stacy et al., 1995) and opiate (Powell et al., 1992) abuse. Therefore, future studies
can employ the IT framework to elucidate the developmental processes underlying hard drug abuse but
should be aware of the types of constructs pertinent to hard drug abuse risk.
5.4.2. Developmental stages and individual differences in substance use outcomes
Although many investigations have identified both mediational and moderational structural pathways,
there is a need to clarify the developmental stages at which each type of pathway is most prevalent. The
two-process framework hypothesizes that expectancies mediate the impact of primary risk factors on use
outcome during early stages of development and moderate their impact during later development. To
investigate this phenomenon, Chorpita and Barlow (1998) suggested that the trend of moderational
effects be modeled over time (i.e., the slope of Risk Expectancy term across developmental stages).
This can be examined by looking at a three-way interaction involving risk, expectancies, and age in
predicting use outcome. If results are consistent with the two-process hypothesis, the three-way
interaction should be the result of a significant moderation of risk by expectancies in older subjects but
no such moderation in younger subjects.
IT pathways might result in different types of outcomes in younger and older individuals. A
preliminary (mediational) process may drive the initiation of use in youths. The escalation and
maintenance of use as well as problems due to drug use may be driven by secondary (moderational)
processes among experienced users at later developmental stages. There is some initial evidence in
support of this notion. One study demonstrated that initiation of substance use in 12-year-olds was
explained by expectancy mediation but not moderation of a protective factor (Trudeau et al., 2003). In a
separate study, the moderating effects of expectancies on family history of alcohol abuse/dependence in
adults were significant only when predicting problem drinking symptoms but not quantity and frequency
of alcohol use (Conway et al., 2003).
5.4.3. Extending the IT model
While the typical expectancy-based IT model solely incorporates three classes of variables, the IT
framework can be extended to examine more specific pathways. For instance, the moderational
pathway can be expanded in order to differentiate between individuals whose use is explained by
moderational processes and those whose use is not explained by such influences. This approach
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2055
attempts to identify a three-way interaction between a primary risk/protective factor, expectancies,
and a secondary factor in the prediction of substance use outcomes. Cooper et al. (1995) utilized
this methodology in demonstrating that the relationship between negative emotions and drinking to
cope was greater among those who held strong positive expectancies and that this moderational
process was strongest in those who were poor copers (a three-way Negative Emotion Expectancy Coping Interaction). Examining three-way interactions such as these might
identify those individuals who may benefit the most of interventions targeting both expectancies and
a primary risk factor.
Mediational pathways in IT models can also be extended. One approach involves demonstrating that
an expectancy-mediated pathway that predicts use, in turn affects a final variable, such as drug-related
social consequences. Kalichman et al. (1998) utilized this method in demonstrating that the influence of
sensation seeking traits on pre-sex substance use in gay and bisexual men was mediated by sexual
enhancement expectancies. They extended this model by showing that pre-sex substance abuse predicted
risky sexual behavior. Mediational models can also be clarified through identification of mediators
between primary expectancies and outcomes. For example, Trudeau et al. (2003) examined an IT model
that identified behavioral willingness to engage in use (i.e., beliefs that one will use if offered drugs) as a
mediator of the effect of alcohol expectancies on alcohol consumption.
5.4.4. Protective influences
Drug expectancy research has not focused much on how expectancies act as protective influences,
except in a few studies (Levy & Earleywine, 2003, 2004; McCarthy, Brown et al., 2001; Trudeau et al.,
2003). Protective factors are thought to have a strong effect on substance abuse outcomes (Chassin,
Carle, Nissim-Sabat, & Kumpfer, 2004); however, little is known about the mechanisms in which they
exert their influence in relation to expectancies (although see Darkes, Greenbaum, & Goldman, 2004;
Galen & Rogers, 2004). Further identification of IT pathways that incorporate primary protective factors
may illuminate potential methods for preventative interventions.
5.4.5. Validating and using the IT model
The clinical relevance of IT models is limited by the degree to which directional causality can be
inferred from IT studies. This is important because if expectancies are merely a consequence or correlate
of addictive behavior rather than causal mediational and moderational factors, they would not be suitable
treatment targets.
Correlational mediational pathways are vulnerable to alternative explanations in which: (1) the
direction of influence among variables might be incorrectly inferred (e.g., substance use might enhance
positive expectancies and/or risk factors); or (2) a fourth (confounding) variable, such as a riskenhancing personality trait or demographic factor, may independently influence all variables, resulting in
their relationships. While prospective correlational studies might avoid the pitfalls of incorrect
directionality because of temporal space between assessments of primary, mediational, and outcome
variables, they remain vulnerable to confounding variable explanations.
Experimental studies that manipulate expectancies and primary risk/protective factors are capable of
demonstrating causality in mediational IT models. To test a specific mediational pathway, two
experiments must be conducted: (i) a first experiment involving manipulation of primary risk/protective
variable; and (ii) a second experiment involving manipulation of the expectancy variable. The
hypothesized mediational pathway is supported if the first experiment demonstrates: (1) significant
2056
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
effects on expectancy and substance use outcome variables in the predicted direction; and (2) statistical
evidence that the effect of the experimental manipulation on substance use was mediated by changes in
expectancies. Validation of the pathway also would require that the second experiment demonstrates that
the manipulation results in: (1) a significant effect of the substance use outcome variable in the predicted
direction; and (2) effects on the primary risk/protective variable that are either non-significant or
significantly smaller than the substance use outcome variable effect. It should be noted that mediational
pathways involving non-malleable primary risk variables (e.g., family history of substance abuse;
demographic characteristics) cannot be tested using this approach.
Associations involving moderation are less vulnerable to plausible alternative explanations (Shadish
et al., 2002). Nonetheless, moderational IT models still can benefit from experimental evaluation. The
experimental approach examines the effect of an intervention targeting expectancies and attempts to
demonstrate a risk factor by treatment interaction in the prediction of substance use outcomes. An
expectancy-based moderational IT pathway is supported if: (1) a risk factor by treatment interaction is
significant; and (2) the interaction is driven by a greater difference between treatment and control
subjects in those that have a primary risk factor (e.g., family history of alcoholism) than those who do
not. A similar logic could be applied to protective factor moderational pathways. Examination of risk/
protection by treatment interactions can identify those individuals who would benefit most from
expectancy challenge interventions (btreatment matchingQ).
A laudable goal of IT pathways research is the identification of as many specific risk trajectories as
possible using methodologies that give the strongest evidence for causal inferences. This literature would
aim to isolate trajectories with specific (as opposed to general) risk/protective factors (e.g., social anxiety
vs. negative affect), expectancy variables (e.g., drinking to enhance social outcomes vs. general positive
outcome expectancies), and substance use outcomes (e.g., alcohol use status vs. severity of alcohol
dependence). In addition, extension of IT models to abuse-related consequences could further add to this
literature.
Identification of specific pathways can inform treatment approaches. For example, one wellresearched IT pathway is the acquired preparedness model (McCarthy, Kroll et al., 2001), which
proposes that the effects of an impulsive personality on substance use is mediated and moderated by
expectancies. Clinicians making use of this information would be aware that the etiological influence of
expectancies is greater in impulsive than non-impulsive patients. Therefore, treatments targeting
expectancies would be strongly suited for such individuals, whereas patients low in impulsivity would
not benefit as greatly from expectancy modification interventions, although they may still be indicated.
Furthermore, clinicians could make use of expectancy mediation. As mentioned in Section 5.2,
mediational pathways indicate that clinicians can treat patients with immutable risk factors by targeting
the expectancies through which they exert their effects. If pathways with specific trajectories are
identified, clinicians can target more specific primary risk factors and expectancy cognitions.
6. Conclusion
There is a growing literature suggesting that drug and alcohol outcome expectancies play a major role
in the development of substance abuse. This paper outlines this research and concludes that expectancies
are a common mediating or moderating etiological process in the development of drug and alcohol
problems. To specify the etiological role of expectancies and other risk factors, a two-process IT model
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2057
of substance abuse risk is presented whereby mediating and moderating pathways involving
expectancies are proposed to occur at different developmental stages. This model points to
expectancy-based targets for prevention and treatment that vary depending on whether preliminary or
secondary processes are operating. Although an initial conceptual model is presented, the developmental
genesis of substance use problems requires further clarification. Future examinations of IT pathways will
likely add to our understanding of the mechanisms underlying substance abuse vulnerability so that more
effective interventions can be developed.
Acknowledgements
This research was supported, in part, by a predoctoral fellowship from the M.D. Anderson Education
Program in Cancer Prevention (R25 CA 57730-11) from the National Cancer Institute awarded to the
first author.
The authors would like to thank Drs. Marc Mooney and Jeremy Pettit for providing valuable
recommendations and comments about this paper.
References
Aarons, G. A., Brown, S. A., Stice, E., & Coe, M. T. (2001). Psychometric evaluation of the marijuana and stimulant effect
expectancy questionnaires for adolescents. Addictive Behaviors, 26(2), 219 – 236.
Adlaf, E. M., & Smart, R. G. (1985). Drug use and religious affiliation, feelings and behaviour. British Journal of Addiction,
80(2), 163 – 171.
Bailey, S. L. (1992). Adolescents’ multisubstance use patterns: the role of heavy alcohol and cigarette use. American Journal of
Public Health, 82(9), 1220 – 1224.
Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological Review, 84(2), 191 – 215.
Barnow, S., Schultz, G., Lucht, M., Ulrich, I., Preuss, U. -W., & Freyberger, H. -J. (2004). Do alcohol expectancies and peer
delinquency/substance use mediate the relationship between impulsivity and drinking behaviour in adolescence? Alcohol
and Alcoholism, 39(3), 213 – 219.
Bridges, L. J., Sigelman, C. K., Brewster, A. B., Leach, D. B., Mack, K. L., Rinehart, C. S., et al. (2003). Cognitive
predictors of children’s attitudes toward alcohol and cocaine. Journal of Child and Adolescent Substance Abuse, 12(3),
19 – 44.
Brown, S. A. (1985a). Expectancies versus background in the prediction of college drinking patterns. Journal of Consulting and
Clinical Psychology, 53(1), 123 – 130.
Brown, S. A. (1985b). Reinforcement expectancies and alcoholism treatment outcome after a one-year follow-up. Journal of
Studies on Alcohol, 46(4), 304 – 308.
Brown, S. A., Christiansen, B. A., & Goldman, M. S. (1987). The alcohol expectancy questionnaire: An instrument for the
assessment of adolescent and adult alcohol expectancies. Journal of Studies on Alcohol, 48(5), 483 – 491.
Brown, S. A., Creamer, V. A., & Stetson, B. A. (1987). Adolescent alcohol expectancies in relation to personal and parental
drinking patterns. Journal of Abnormal Psychology, 96(2), 117 – 121.
Brown, S. A., Goldman, M. S., Inn, A., & Anderson, L. R. (1980). Expectations of reinforcement from alcohol: Their domain
and relation to drinking patterns. Journal of Consulting and Clinical Psychology, 48(4), 419 – 426.
Brown, S. A., Tate, S. R., Vik, P. W., Haas, A. L., & Aarons, G. A. (1999). Modeling of alcohol use mediates the effect of
family history of alcoholism on adolescent alcohol expectancies. Experimental and Clinical Psychopharmacology, 7(1),
20 – 27.
Casswell, S., Gilmore, L. L., Silva, P., & Brasch, P. (1988). What children know about alcohol and how they know it. British
Journal of Addiction, 83(2), 223 – 227.
2058
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
Chassin, L., Carle, A. C., Nissim-Sabat, D., & Kumpfer, K. L. (2004). Fostering resilience in children of alcoholic parents.
In K. I. Maton, & C. J. Schellenbach (Eds.), Investing in children, youth, families, and communities: Strengths-based
research and policy (pp. 137 – 155). American Psychological Association.
Chorpita, B. F., & Barlow, D. H. (1998). The development of anxiety: The role of control in the early environment.
Psychological Bulletin, 124(1), 3 – 21.
Cloninger, C. R. (1987). A systematic method for clinical description and classification of personality variants. A proposal.
Archives of General Psychiatry, 44(6), 573 – 588.
Cloninger, C. R., Sigvardsson, S., & Bohman, M. (1988). Childhood personality predicts alcohol abuse in young adults.
Alcoholism, Clinical and Experimental Research, 12(4), 494 – 505.
Cohen, P., Brown, J., & Smailes, E. (2001). Child abuse and neglect and the development of mental disorders in the general
population. Development and Psychopathology, 13(4), 981 – 999.
Cole, D. A., & Turner, J. E. (1993). Models of cognitive mediation and moderation in child depression. Journal of Abnormal
Psychology, 102(2), 271 – 281.
Compton, W. M., Cottler, L. B., Ridenour, T., Ben-Abdallah, A., & Spitznagel, E. L. (2002). The specificity of family history of
alcohol and drug abuse in cocaine abusers. American Journal of Addiction, 11(2), 85 – 94.
Conley, C. S., Haines, B. A., Hilt, L. M., & Metalsky, G. I. (2001). The children’s attributional style interview:
Developmental tests of cognitive diathesis–stress theories of depression. Journal of Abnormal Child Psychology, 29(5),
445 – 463.
Conway, K. P., Swendsen, J. D., & Merikangas, K. R. (2003). Alcohol expectancies, alcohol consumption, and problem
drinking: The moderating role of family history. Addictive Behaviors, 28(5), 823 – 836.
Cooper, M. L., Frone, M. R., Russell, M., & Mudar, P. (1995). Drinking to regulate positive and negative emotions: A
motivational model of alcohol use. Journal of Personality and Social Psychology, 69(5), 990 – 1005.
Cooper, M. L., Russell, M., Skinner, J. B., Frone, M. R., & Mudar, P. (1992). Stress and alcohol use: Moderating effects of
gender, coping, and alcohol expectancies. Journal of Abnormal Psychology, 101(1), 139 – 152.
Copeland, A. L., & Carney, C. E. (2003). Smoking expectancies as mediators between dietary restraint and disinhibition and
smoking in college women. Experimental and Clinical Psychopharmacology, 11(3), 247 – 251.
Cox, W. M., Hogan, L. M., Kristian, M. R., & Race, J. H. (2002). Alcohol attentional bias as a predictor of alcohol abusers’
treatment outcome. Drug and Alcohol Dependence, 68(3), 237 – 243.
Crum, R. M., Ensminger, M. E., Ro, M. J., & McCord, J. (1998). The association of educational achievement and school
dropout with risk of alcoholism: A twenty-five-year prospective study of inner-city children. Journal of Studies on Alcohol,
59(3), 318 – 326.
Cruz, I. Y., & Dunn, M. E. (2003). Lowering risk for early alcohol use by challenging alcohol expectancies in elementary school
children. Journal of Consulting and Clinical Psychology, 71(3), 493 – 503.
D’Amico, E. J., & Fromme, K. (2000). Implementation of the risk Skills training program: A brief intervention targeting
adolescent participation in risk behaviors. Cognitive and Behavioral Practice, 7(1), 101 – 117.
Darkes, J., & Goldman, M. S. (1993). Expectancy challenge and drinking reduction: Experimental evidence for a mediational
process. Journal of Consulting and Clinical Psychology, 61(2), 344 – 353.
Darkes, J., & Goldman, M. S. (1998). Expectancy challenge and drinking reduction: Process and structure in the alcohol
expectancy network. Experimental and Clinical Psychopharmacology, 6(1), 64 – 76.
Darkes, J., Greenbaum, P. E., & Goldman, M. S. (2004). Alcohol expectancy mediation of biopsychosocial risk: Complex
patterns of mediation. Experimental and Clinical Psychopharmacology, 12(1), 27 – 38.
Dunn, M. E., & Goldman, M. S. (1998). Age and drinking-related differences in the memory organization of
alcohol expectances in 3rd-, 6th-, 9th-, and 12th-grade children. Journal of Consulting and Clinical Psychology, 66(3),
579 – 585.
Dunn, M. E., Lau, H. C., & Cruz, I. Y. (2000). Changes in activation of alcohol expectancies in memory in relation to changes
in alcohol use after participation in an expectancy challenge program. Experimental and Clinical Psychopharmacology, 8(4),
566 – 575.
Fearnow-Kenny, M. D., Wyrick, D. L., Hansen, W. B., Dyreg, D., & Beau, D. B. (2001). Normative beliefs, expectancies, and
alcohol-related problems among college students: Implications for theory and practice. Journal of Alcohol and Drug
Education, 47(1), 31 – 44.
Felix-Ortiz, M., & Newcomb, M. D. (1999). Vulnerability for drug use among Latino adolescents. Journal of Community
Psychology, 27(3), 257 – 280.
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2059
Frank, S. J., Jacobson, S., & Tuer, M. (1990). Psychological predictors of young adults’ drinking behaviors. Journal of
Personality and Social Psychology, 59(4), 770 – 780.
Furr-Holden, C. D. M., & Anthony, J. C. (2003). Epidemiologic differences in drug dependence: A US–UK cross-national
comparison. Social Psychiatry and Psychiatric Epidemiology, 38(4), 165 – 172.
Galen, L. W., & Rogers, W. M. (2004). Religiosity, alcohol expectancies, drinking motives and their interaction in the
prediction of drinking among college students. Journal of Studies on Alcohol, 65(4), 469 – 476.
Goldman, M. S., & Darkes, J. (2004). Alcohol expectancy multiaxial assessment: A memory network-based approach.
Psychological Assessment, 16(1), 4 – 15.
Gray, J. A. (1991). The neuropsychology of temperament. In J. Strelau, & A. Angleitner (Eds.), Explorations in temperament:
International perspectives on theory and measurement. Perspectives on individual differences (pp. 105 128).
Griffin, K. W., Epstein, J. A., Botvin, G. J., & Spoth, R. L. (2001). Social competence and substance use among rural youth:
Mediating role of social benefit expectancies of use. Journal of Youth and Adolescence, 30(4), 485 – 498.
Hartzler, B., & Fromme, K. (2003). Cognitive–behavioral profiles of college risk-takers with Type II and psychopathic
personality traits. Addictive Behaviors, 28(2), 315 – 326.
Hine, D. W., McKenzie-Richer, A., Lewko, J., Tilleczek, K., & Perreault, L. (2002). A comparison of the mediational properties
of four adolescent smoking expectancy measures. Psychology of Addictive Behaviors, 16(3), 187 – 195.
Inhelder, B., & Piaget, J. (1969). The early growth of logic in the child. New York7 Norton.
Johnston, L. D., & O’Malley, P. M. B. J. G. (1986). Drug use among American high school students, college students, and other
young adults: National trends through 1985. MD7 National Institute on Drug Abuse: Rockville.
Johnston, L. D., O’Malley, P. M., & Bachman, J. G. (1987). Psychotherapeutic, licit, and illicit use of drugs among adolescents.
An epidemiological perspective. J Adolesc Health Care, 8(1), 36 – 51.
Johnstone, B. (1994). Sociodemographic, environmental, and cultural influences on adolescent drinking behavior. In R. Zucker,
G. Boyd, & J. Howard (Eds.), The development of alcohol problems: Exploring the biopsychosocial matrix of risk. Research
Monograph, vol. 26 (pp. 169 203). Rockville, MD7 National Institute on Alcohol Abuse and Alcoholism.
Jones, B. T., Corbin, W., & Fromme, K. (2001). A review of expectancy theory and alcohol consumption. Addiction, 96(1),
57 – 72.
Kalichman, S. C., Tannenbaum, L., & Nachimson, D. (1998). Personality and cognitive factors influencing substance use and
sexual risk for HIV infection among gay and bisexual men. Psychology of Addictive Behaviors, 12(4), 262 – 271.
Katz, E. C., Fromme, K., & D’Amico, E. J. (2000). Effects of outcome expectancies and personality on young adults’ illicit
drug use, heavy drinking, and risky sexual behavior Erratum. Cognitive Therapy and Research, 24(6), 781.
Kidorf, M., & Lang, A. R. (1999). Effects of social anxiety and alcohol expectancies on stress-induced drinking. Psychology of
Addictive Behaviors, 13(2), 134 – 142.
Kilpatrick, D. G., Acierno, R., Saunders, B., Resnick, H. S., Best, C. L., & Schnurr, P. P. (2000). Risk factors for adolescent
substance abuse and dependence: Data from a national sample. Journal of Consulting and Clinical Psychology, 68(1), 19 – 30.
Kivlahan, D. R., Marlatt, G. A., Fromme, K., & Coppel, D. B. (1990). Secondary prevention with college drinkers: Evaluation
of an alcohol skills training program. Journal of Consulting and Clinical Psychology, 58(6), 805 – 810.
Ladd, G. W., & Burgess, K. B. (2001). Do relational risks and protective factors moderate the linkages between childhood
aggression and early psychological and school adjustment? Child Development, 72(5), 1579 – 1601.
Leigh, B. C. (1987). Evaluations of alcohol expectancies: Do they add to prediction of drinking patterns? Psychology of
Addictive Behaviors, 1(3), 135 – 139.
Leigh, B. C. (1989a). Confirmatory factor analysis of alcohol expectancy scales. Journal of Studies on Alcohol, 50(3), 268 – 277.
Leigh, B. C. (1989b). In search of the seven dwarves: Issues of measurement and meaning in alcohol expectancy research.
Psychological Bulletin, 105(3), 361 – 373.
Leigh, B. C., & Stacy, A. W. (1993). Alcohol outcome expectancies: Scale construction and predictive utility in higher order
confirmatory models. Psychological Assessment, 5(2), 216 – 229.
Levy, B., & Earleywine, M. (2003). Reinforcement expectancies for studying predict drinking problems among college
students: Approaching drinking from an expectancies choice perspectives. Addictive Behaviors, 28(3), 551 – 559.
Levy, B., & Earleywine, M. (2004). Discriminating reinforcement expectancies for studying from future time perspective in the
prediction of drinking problems. Addictive Behaviors, 29(1), 181 – 190.
Lewis, M. (2000). Toward a development of psychopathology: Models, definitions, and prediction. In Arnold J. Sameroff,
Lewis Michael, et al., (Eds.), Handbook of developmental psychopathology (2nd ed.). Dordrecht, Netherlands7 Kluwer
Academic Publishers.
2060
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
Luthar, S. S., Cicchetti, D., & Becker, B. (2000). The construct of resilience: A critical evaluation and guidelines for future
work. Child Development, 71(3), 543 – 562.
Margolin, G., & Gordis, E. B. (2000). The effects of family and community violence on children. Annual Review of Psychology,
51, 445 – 479.
Marlatt, G. A., Baer, J. S., Kivlahan, D. R., Dimeff, L. A., Larimer, M. E., Quigley, L. A., et al. (1998). Screening and brief
intervention for high-risk college student drinkers: Results from a 2-year follow-up assessment. Journal of Consulting and
Clinical Psychology, 66(4), 604 – 615.
Maughan, A., & Cicchetti, D. (2002). Impact of child maltreatment and interadult violence on children’s emotion regulation
abilities and socioemotional adjustment. Child Development, 73(5), 1525 – 1542.
McCarthy, D. M., Aarons, G. A., & Brown, S. A. (2002). Educational and occupational attainment and drinking behavior: An
expectancy model in young adulthood. Addiction, 97(6), 717 – 726.
McCarthy, D. M., Brown, S. A., Carr, L. G., & Wall, T. L. (2001). ALDH2 status, alcohol expectancies, and alcohol response:
Preliminary evidence for a mediation model. Alcoholism: Clinical and Experimental Research, 25(11), 1558 – 1563.
McCarthy, D. M., Kroll, L. S., & Smith, G. T. (2001). Integrating disinhibition and learning risk for alcohol use. Experimental
and Clinical Psychopharmacology, 9(4), 389 – 398.
McCarthy, D. M., Wall, T. L., Brown, S. A., & Carr, L. G. (2000). Integrating biological and behavioral factors in alcohol use
risk: The role of ALDH2 status and alcohol expectancies in a sample of Asian Americans. Experimental and Clinical
Psychopharmacology, 8(2), 168 – 175.
McCusker, C. G. (2001). Cognitive biases and addiction: An evolution in theory and method. Addiction, 96(1), 47 – 56.
Merikangas, K. R. (2002). Implications of genetic epidemiology for classification. In J. E. Helzer, & J. J. Hudziak
(Eds.), Defining psychopathology in the 21st century: DSM-V and beyond (pp. 195 – 209). American Psychiatric Publishing,
Inc..
Miller, P. M., Smith, G. T., & Goldman, M. S. (1990). Emergence of alcohol expectancies in childhood: A possible critical
period. Journal of Studies on Alcohol, 51(4), 343 – 349.
Moeller, F. G., & Dougherty, D. M. (2002). Impulsivity and substance abuse: What is the connection? Addictive Disorders and
Their Treatment, 1(1), 3 – 10.
Mooney, D. K., & Corcoran, K. J. (1989). The relationship between assertiveness, alcohol-related expectations for social
assertion and drinking patterns among college students. Addictive Behaviors, 14(3), 301 – 305.
Moore, J. (1994). The chola life course: Chicana heroin users and the barrio gang. International Journal of the Addictions,
29(9), 1115 – 1126.
Moss, H. B., Lynch, K. G., & Hardie, T. L. (2003). Affiliation with deviant peers among children of substance dependent
fathers from pre-adolescence into adolescence: Associations with problem behaviors. Drug and Alcohol Dependence, 71(2),
117 – 125.
Musher-Eizenman, D. R., & Kulick, A. D. (2003). An alcohol expectancy-challenge prevention program for at risk college
women. Psychology of Addictive Behaviors, 17(2), 163 – 166.
Oei, T. P. S., & Morawska, A. (2004). A cognitive model of binge drinking: The influence of alcohol expectancies and drinking
refusal self-efficacy. Addictive Behaviors, 29(1), 159 – 179.
Ohannessian, C. M., Hesselbrock, V. M., Tennen, H., & Affleck, G. (1994). Hassles and uplifts and generalized outcome
expectancies as moderators on the relation between a family history of alcoholism and drinking behaviors. Journal of
Studies on Alcohol, 55(6), 754 – 763.
Ouellette, J. A., Gerrard, M., Gibbons, F. X., & Reis-Bergan, M. (1999). Parents, peers, and prototypes: Antecedents of
adolescent alcohol expectancies, alcohol consumption, and alcohol-related life problems in rural youth. Psychology of
Addictive Behaviors, 13(3), 183 – 197.
Patterson, C. M., & Newman, J. P. (1993). Reflectivity and learning from aversive events: Toward a psychological mechanism
for the syndromes of disinhibition. Psychological Review, 100(4), 716 – 736.
Piaget, J. (1962). Play, dreams, and imitation in childhood. New York7 Norton.
Powell, J., Bradley, B., & Gray, J. (1992). Classical conditioning and cognitive determinants of subjective craving for opiates:
An investigation of their relative contributions. British Journal of Addiction, 87(8), 1133 – 1144.
Rather, B. C., & Goldman, M. S. (1994). Drinking-related differences in the memory organization of alcohol expectancies.
Experimental and Clinical Psychopharmacology, 2(2), 167 – 183.
Roehrich, L., & Goldman, M. S. (1995). Implicit priming of alcohol expectancy memory processes and subsequent drinking
behavior. Experimental and Clinical Psychopharmacology, 3(4), 402 – 410.
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
2061
Schafer, J., & Brown, S. A. (1991). Marijuana and cocaine effect expectancies and drug use patterns. Journal of Consulting and
Clinical Psychology, 59(4), 558 – 565.
Schuckit, M. A. (1994). A clinical model of genetic influences in alcohol dependence. Journal of Studies on Alcohol, 55(1),
5 – 17.
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal
inference. Houghton Mifflin Co.
Sher, K. J., Walitzer, K. S., Wood, P. K., & Brent, E. E. (1991). Characteristics of children of alcoholics: Putative risk factors,
substance use and abuse, and psychopathology. Journal of Abnormal Psychology, 100(4), 427 – 448.
Sher, K. J., Wood, M. D., Wood, P. K., & Raskin, G. (1996). Alcohol outcome expectancies and alcohol use: A latent variable
cross-lagged panel study. Journal of Abnormal Psychology, 105(4), 561 – 574.
Simons-Morton, B. (2004). Prospective association of peer influence, school engagement, drinking expectancies, and parent
expectations with drinking initiation among sixth graders. Addictive Behaviors, 29(2), 299 – 309.
Smith, G. T., Goldman, M. S., Greenbaum, P. E., & Christiansen, B. A. (1995). Expectancy for social facilitation from
drinking: The divergent paths of high-expectancy and low-expectancy adolescents. Journal of Abnormal Psychology,
104(1), 32 – 40.
Spooner, C. (1999). Causes and correlates of adolescent drug abuse and implications for treatment. Drug and Alcohol Review,
18(4), 453 – 475.
Stacy, A. W., & Newcomb, M. D. (1998). Memory association and personality as predictors of alcohol use: Mediation and
moderator effects. Experimental and Clinical Psychopharmacology, 6(3), 280 – 291.
Stacy, A. W., Newcomb, M. D., & Bentler, P. M. (1991). Cognitive motivation and drug use: A 9-year longitudinal study.
Journal of Abnormal Psychology, 100(4), 502 – 515.
Stacy, A. W., Newcomb, M. D., & Bentler, P. M. (1995). Expectancy in mediational models of cocaine use. Personality and
Individual Differences, vol. 19 (pp. 655 – 667). Elsevier Science.
Stipek, D., Recchia, S., & McClintic, S. (1992). Self-evaluation in young children. Monographs of the Society for Research in
Child Development, 57(1), 100 (Mono 226).
Sussman, S., Dent, C. W., & Stacy, A. W. (1999). The association of current stimulant use with demographic,
substance use, violence-related, social and intrapersonal variables among high risk youth. Addictive Behaviors, 24(6),
741 – 748.
Tapert, S. F., McCarthy, D. M., Aarons, G. A., Schweinsburg, A. D., & Brown, S. A. (2003). Influence of language
abilities and alcohol expectancies on the persistence of heavy drinking in youth. Journal of Studies on Alcohol, 64(3),
313 – 321.
Trudeau, L., Lillehoj, C., Spoth, R., & Redmond, C. (2003). The role of assertiveness and decision making in early adolescent
substance initiation: Mediating processes. Journal of Research on Adolescence, 13(3), 301 – 328.
Turner, J. E., Jr., & Cole, D. A. (1994). Developmental differences in cognitive diatheses for child depression. Journal of
Abnormal Child Psychology, 22, 15 – 32.
Uhl, G., Blum, K., Noble, E., & Smith, S. (1993). Substance abuse vulnerability and D2 receptor genes. Trends in
Neurosciences, 16(3), 83 – 88.
Van Hasselt, V. B., Null, J. A., Kempton, T., & Bukstein, O. G. (1993). Social skills and depression in adolescent substance
abusers. Addictive Behaviors, 18(1), 9 – 18.
Vangsness, L., Bry, B. H., & LaBouvie, E. W. (2005). Impulsivity, negative expectancies, and marijuana use: A test of the
acquired preparedness model. Addictive Behaviors, 30(5), 1071 – 1076.
Voelkl, K. E., & Frone, M. R. (2000). Predictors of substance use at school among high school students. Journal of Educational
Psychology, 92(3), 583 – 592.
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA7 Harvard
University Press.
Vygotsky, L. S. (1986). Thought and language. Cambridge, MA7 MIT Press.
Wall, T. L., Thomasson, H. R., Schuckit, M. A., & Ehlers, C. L. (1992). Subjective feelings of alcohol intoxication in Asians
with genetic variations of ALDH2 alleles. Alcoholism: Clinical and Experimental Research, 16(5), 991 – 995.
Waters, A. J., Shiffman, S., Sayette, M. A., Paty, J. A., Gwaltney, C. J., & Balabanis, M. H. (2003). Attentional bias predicts
outcome in smoking cessation. Health Psychology, 22(4), 378 – 387.
Weingardt, K. R., Stacy, A. W., & Leigh, B. C. (1996). Automatic activation of alcohol concepts in response to positive
outcomes of alcohol use. Alcoholism: Clinical and Experimental Research, 20(1), 25 – 30.
2062
A.M. Leventhal, J.M. Schmitz / Addictive Behaviors 31 (2006) 2038–2062
Wiers, R. W., Van Woerden, N., Smulders, F. T. Y., & De Jong, P. J. (2002). Implicit and explicit alcohol-related cognitions in
heavy and light drinkers. Journal of Abnormal Psychology, 111(4), 648 – 658.
Willner, P. (2001). A view through the gateway: Expectancies as a possible pathway from alcohol to cannabis. Addiction, 96(5),
691 – 703.
Wood, M. D., Read, J. P., Palfai, T. P., & Stevenson, J. F. (2001). Social influence processes and college student drinking: The
mediational role of alcohol outcome expectations. Journal of Studies on Alcohol, 62(1), 32 – 43.
Yu, J., & Williford, W. R. (1992). The age of alcohol onset and alcohol, cigarette, and marijuana use patterns: An analysis of
drug use progression of young adults in New York State. International Journal of the Addictions, 27(11), 1313 – 1323.
Zimbardo, P. G., & Boyd, J. N. (1999). Putting time in perspective: A valid, reliable individual-differences metric. Journal of
Personality and Social Psychology, 77(6), 1271 – 1288.
Zogg, J. B., Ma, H., Dent, C. W., & Stacy, A. W. (2004). Self-generated alcohol outcomes in 8th and 10th graders: Exposure to
vicarious sources of alcohol information. Addictive Behaviors, 29(1), 3 – 16.