Alcohol and Drug Use and Self-Esteem:
A Psychological Perspective

Rodney Skager and Elizabeth Kerst

Introduction

In the fall of 1987, a network television commercial gave us a putative insider's view of creative thinking in the auto industry. The setting was a meeting of attractive young executives struggling to define what people wanted from their automobiles. The group was casually arranged around a conference table. Various signs suggested that it was near the end of the working day: the men's ties were loose, while the women sported unruly wisps of hair.

Each participant haltingly tried to capture the essence of what people were looking for. One attempt after another fell short. Finally, one of the young men hit pay dirt. "This may sound like a cliché," he said, "but you are what you drive." Here was our message. We are what we drive.

Apt as it may be for our times, the commercial is merely an example of how we define ourselves by things in our environment, especially possessions visible to other people. According to this view, who we think we are and how we feel about who we are are based on what we believe others think of us. But there is also an opposite conception, one that internalizes identity and its associated sense of self-worth. Within this perspective, the person who derives his or her sense of self primarily from the reactions of others is unlikely to achieve a firm sense of identity. If it is really a central structural element of personality, personal

identity must be stable and self-maintaining in the mature person. Otherwise, how are we to account for the autonomous human being? Autonomous people can endure being ignored or even devalued by others, because they are supported by a strong inner sense of self that sustains independent thought and action.

Self-esteem is the experience of one's personal self-worth. It is a mental indicator, as body temperature and blood pressure are physical indicators. But the analogy has limits. Self-esteem tends to be experienced as either present or absent, rather than in precise gradations. Some people, especially those who are abnormally dependent on others for praise and recognition, may experience significant shifts in self-worth, depending on how much support they receive from other people. Although such individuals may be described as lacking self-esteem, it would be more accurate to characterize them as having a weak sense of self, which is revealed through shifts in self-esteem. People of this type often need constant admiration and validation from others. They cannot get enough of what some observers like to call "narcissistic supplies," literally, fuel for self-love.

When their self-esteem is very low, some people are depressed and apathetic, sometimes even suicidal. Others repeatedly form nonsupportive relationships in which they are devalued and demeaned. Suicidal behavior and the need for self-destructive relationships suggest serious problems with self-concept. But so do narcissistic disorders in which people appear to have very high levels of self-esteem, while behaving with arrogance and disregard for others. Unhealthy self-esteem may also be associated with grandiose and delusional thinking among people suffering from psychoses. This kind of self-esteem can be described as the pathological result of a failed attempt to develop a stable, cohesive self or identity.

Pathological conditions result from defects in personality development occurring early in life. People who suffer from such conditions attempt to maintain self-esteem through processes that are different from, or are serious distortions of, those used by people who developed in a relatively normal way. Healthy people do seek out and maintain social contacts that support or enhance their sense of self-worth. But they do not require constant praise and recognition from others, and they are uncomfortable about unearned praise.

In people who have achieved extraordinary levels of personal development, high self-esteem and humility coexist. Such people have a strong, enduring sense of personal identity, accepting in themselves the

personal inadequacies and failures that exist in all of us. Most important, their strong, but realistic, sense of self does not have to be sustained at the expense of others. They do not need to control or humiliate other people. Rather, their strong identity allows them to treat others with fairness and consideration.

Psychological theories usually assume that personality structures maintaining self-esteem originate early in life. These theories also allow for further development in social and other situations encountered throughout life. For example, in Mahler, Pine, and Bergman's (1975) or Kohut's (1977) theories, structures that maintain self-esteem are gradually internalized through a process of appropriate parenting.

There is no doubt that self-esteem is central in the consciousness of troubled human beings. Psychotherapists report that those who seek help typically suffer from low self-esteem. Jerome Frank has characterized the mental state of those who seek psychotherapy as one of personal demoralization: "A plausible hypothesis is that patients seek psychotherapy not for symptoms alone, but for symptoms coupled with demoralization: subjective incompetence, loss of self-esteem, alienation, hopelessness (feeling that no one can help), or helplessness (feeling that other people could help, but will not)" (1982, 16).

This description applies equally well to the alcoholic. The book Alcoholics Anonymous, for example, describes the mental state of the alcoholic seeking help as a state of "pitiful and incomprehensible demoralization" (1976, 30). The first step of the Alcoholics Anonymous (AA) program of action asks recovering persons to accept the proposition, "We were powerless over alcohol; our lives had become unmanageable."

The themes of "powerlessness" and "unmanageability" recur in the literature of AA and other support groups modeled on its program. Members of AA believe that the alcoholic or other addict finally turns to others for assistance when denial and rationalization give way to a profound sense of helplessness in the face of a life that has spun out of control. AA describes this crisis as "hitting bottom." More than anything else, hitting bottom is a state of negative self-worth, a vacuum where self-denigration replaces self-esteem.

Theories of Self-Concept and Identity

If self-esteem is an indicator or sign of the quality of an internalized structure we call "the self," then it needs to be related to behavior through a theory about personality. This anticipates the basic assump-

tion of this chapter: the relationship between substance use and self-esteem is not a simple causal relationship but, rather, one that is mediated through an organized aspect of the personality we choose to call "self-concept" or "identity."

Self-Concept and Self-Esteem

Markus and Wurf (1987) have recently contributed a comprehensive review of research on self-concept . They make an observation that, from the perspective of a book on self-esteem, is somewhat disquieting: "The majority of self-concept research could best be described as an attempt to relate very complex global behavior, such as delinquency, marital satisfaction, or school achievement, to a single aspect of the self-concept, typically self-esteem" (1987, 300).

In other words, the self-concept is now understood not only to incorporate self-esteem, but also to have considerably broader meaning. Markus and Wurf show that research on self-concept over the past decade has progressed beyond studying self-esteem to an emphasis on self-concept as a "dynamic interpretive structure," which mediates both intrapersonal processes such as information processing, affect, and motivation and interpersonal processes such as choice of social partner and situation, interaction strategy, and reaction to evaluations from others.

Marsh and Shavelson (1985) have demonstrated that self-concept is multidimensional. That is, it is composed of various self-representations that differ from one another in importance and, according to Markus and Wurf (1987), even in whether or not they have been achieved. Unachieved selves, termed possible selves, may be desirable or undesirable. They function as incentives for behavior in the sense of being end states to achieve or avoid. Higgins (1983) extends this conception by suggesting three classes of self-representations: the actual self; the ideal self, or self that the individual would like to achieve; and the ought self, or self that an individual or others think one should achieve. Discrepancies between actual and ideal self-conceptions are associated with depression, those between actual and ought associated with anxiety. Alternatively, self-conceptions may be divided into good and bad. The latter are readily identified with depression (Sullivan 1953; Beck 1967), which is a clinically defined state characterized by abnormally low self-esteem.

According to Markus and Wurf, identity is "an image of the self that one tries to convey to others" (1987, 325). It is thus both a self-

conception and an "entity out in the world." Depending on their goals and the audience, people try to construct different identities. This is an important way in which self-representations influence human behavior. Self-representations are not always directly manifest in overt actions, however; they are often seen indirectly in "mood changes, in variations in what aspects of the self-concept are accessible and dominant, in shifts in self-esteem, in social comparison choices, in the nature of self-presentation, in choice of social setting, and in the construction and definition of one's situation" (Markus and Wurf 1987, 300). These observations are consistent with our view that self-esteem is an experiential state that varies as a function of an underlying self-representation within a situational context. The goals of the individual within a situation, as well as the individual's life history, mediate the relationship.

None of this denies the significance of self-esteem, as long as it is understood as a state that reflects the quality or adequacy of a self-representation. The self-representation takes on an organizing function in the personality. It accounts for how an individual interprets a situation in personal terms, and it focuses, organizes, and directs behavior in that situation.

The recovering alcoholic develops an explicit self-representation that is also an identity (a public self-concept). When speaking at an AA meeting, he or she learns to say, "I am an alcoholic." During recovery, this initially negative self-representation is transformed, becoming to its holder a positive self-representation incorporating the ideas of sobriety and recovery. The new identity incorporates a set of organizing principles for living. Sober alcoholics are people who do not take the first drink, who stay out of "slippery places," who value abstinence (unlike the society they live in), who reinforce their own sobriety by helping other alcoholics, and who have achieved a sense of place and belonging in a supportive community. These achievements transform self-derogation into self-esteem through the creation of a substantial, positive self-representation.

This example suggests that efforts to replenish self-esteem without regard to developing a healthy self-representation or identity are mis-directed. People who continually seek praise and adulation to bolster their self-worth are only temporarily satisfied. They desperately need instead to engage in a process of personal growth that allows them to internalize their sense of self, so that they can feel self-worth without continuous bolstering from others or from alcohol and drugs.

In order to understand the relationship between self-esteem and sub-

stance use, a theory about how self-esteem is generated and maintained in the personality is needed. To simply ascertain that people drink or use drugs excessively because they have low self-esteem tells us nothing about prevention and remediation unless we also understand the origins of self-esteem.

The Self in Object Relations Theory

Object relations theories have developed within psychodynamic psychology as alternatives to the more mechanistic formulations of psychoanalytic theory, especially the Freudian concept of drives. In contrast to the latter, with its emphasis on internal sexual and aggressive forces and the mechanisms that control them, object relations theory is concerned with how parents and other caretakers shape the psychological development of the infant and child.

"Objects" are usually, but not always, other people. For example, the most celebrated "object" may be Linus's blanket in the "Peanuts" comic strip. In object relations theory, the blanket and the class of childhood objects it represents are referred to as "transitional objects," which smooth the transition from dependence on the soothing function of an adult, ordinarily the mother, to a state of separation and individuation in which the child becomes increasingly self-sufficient emotionally. The child uses the soothing provided by the transitional object to replace that originally provided by the parent. The behavior of children in relation to their transitional objects has been described by one theorist as very much like the behavior of addicted adults in relation to the "object" of their addiction (Tolpin 1971).

Although many theorists and researchers have played significant roles in the development of object relations theory, the contributions of Heinz Kohut and Margaret Mahler are most relevant to abusive or addictive use of alcohol and other drugs. Kohut (1977) conceives of the development of the self in terms of relationships with early "self-objects," especially the mother. Self-objects are other people over whom an infant (or adult) feels a sense of control . According to Kohut, this sense of control has a special quality, resembling the control experienced over one's own self. There is a merged quality in the relationship; the infant does not perceive a self-object as a separate person.

The idea that an infant "controls" an adult may seem farfetched, but it is based on sound clinical observation and reasoning. Adequately cared-for infants do control the adults around them, at least with re-

spect to having their personal needs met. The infant's sense of controlling the parent is accurate within its own perspective. As maturation proceeds and the developing child's needs and capabilities become more complex, the sense of control over the self-object is gradually lost, at least in normal development. The process of separating and individuating is the major developmental task of early childhood.

Kohut's theoretical work evolved around the theme of the development of the self in relation to early self-objects. For example, maternal and other adult self-objects model organizing and soothing functions, which are gradually internalized by the child ("the hurt will go away in a minute").

When parents and others give recognition and praise to a very young child, they are engaging in a process called "mirroring." Their mirroring helps the child define an early self or identity. Under conditions of adequate parenting, this early self is grandiose, even omnipotent. Wise parents do not criticize toddlers or arrange experiences of failure in an attempt to teach extremely young children to develop a realistic view of their own capabilities. Mahler, Pine, and Bergman, who conducted long-term observational studies of mothers and children, characterize this period of development as the "practicing subphase" of a much longer process of separation from the primary parental self-object and individuation through achievements, "marking the child's assumption of his own individual characteristics" (1975, 4). The essential scenario of this practicing subphase casts the toddler as the recipient of recognition and praise for every act of individual achievement—hence the grandiosity of the earliest self-structure, according to the object relations theorists.

If the self is a grandiose construction, then self-esteem must be high, as long as other people are supportive. But this kind of self-esteem does not have a stable base. It is easily upset by frustration and failure, which soon occur in the life of every normally developing child. In Mahler's developmental scheme, the practicing subphase is associated with the period of crawling and early walking (from ten or twelve months to sixteen or eighteen months) and is followed by a "rapprochement phase," in which the child begins to perceive the reality of being small and inept. If, for various reasons discussed by object relations theorists, the child is prevented from negotiating this next phase of development, he or she will fail to develop a more mature self-organization or will at least retain elements of grandiosity and omnipotence that deter or distort later development.

This failure is the origin of the narcissistic personality, which Kohut and Wolf (1978) and others have associated with addictive behavior. Narcissistic individuals often show an inflated sense of self-worth, but one that is unstable and not self-sustaining, that is, it is highly dependent on a sustaining environment for its replenishment. Narcissistic people are exceptionally sensitive to failure, criticism, and being ignored or slighted. They also tend to relate to others primarily in terms of how well those others contribute to the satisfaction of personal needs. A narcissistic personality involves a sense of "entitlement," under which other people may be used and then discarded when they no longer prove useful. It is an abnormal or unrealistic self-representation (and associated self-esteem) that mediates such callous, insensitive behavior and disregard of the rights of others.

People who successfully negotiate later developmental phases develop a self that is stable and, in Kohut's terms, "firm and coherent." They can tolerate criticism, failure, and devaluation by others, because they possess a self-structure that remains constant even in situations of devaluation or failure. In this process of development, they were assisted by parents who not only gave support when needed but who also knew when to stand aside and let children learn to deal with life's problems. In Bandura's (1984) terms, they maintained conditions under which a child could develop a sense of self-efficacy, one of the conditions important to the development of a healthy self. Existential psychologists refer to a similar idea in the concept of "hardiness" (Kobasa 1979), which they see as a precondition to achieving autonomy as an adult.

Kohut (1977) concluded that a firm and coherent self fails to develop when a child does not achieve either an initial mirroring relationship with a parental self-object (usually the mother) or a later identification with an idealized parental figure (either the mother or the father). Chelton and Bonney (1987) suggest that an addiction reactivates emotions associated with the developmental stage at which a critical failure in object relationship occurred. The addictive use of alcohol or drugs may be viewed as an abortive attempt to recreate a primitive mental state, from which interrupted growth can begin anew. Addiction, of course, is a flawed solution.

Ego States and the Fragmented Self

Kohut's (1977) descriptions of the fragmented or noncohesive self are paralleled in the work of object relations theorists using a more tradi-

tional terminology. Kernberg (1966) helped lay a theoretical foundation for the observation that changes in personality often occur in people who are under the influence of alcohol or other drugs. For example, even when they are only moderately drunk, quiet and shy people may become outgoing and daring in social interactions, the meek may become assertive, the asexual sexual, and so on. Occasionally, complete changes of identity occur in alcoholic "blackouts," dissociative states that are walled off from ordinary consciousness by forgetting (amnesia).

These phenomena can be interpreted as changes in self or identity. As such, they represent alternative "ego states," or, in Kernberg's terms, "compartmentalized psychic manifestations" (1966, 236). Such states originate in response to inconsistencies in parental behavior that remain unresolved during later development. Primitive self-organizations arise in response to such inconsistencies, each corresponding to representations of the parental object, the associated "part-self," and emotional states associated with each. For example, a child with an alcoholic parent usually experiences dramatic changes in that parent associated with whether he or she is drunk or sober. Alternative selves develop accordingly, with the two part-selves defensively separated, so that feelings associated with one parental self-object do not spill over into the other. The "good parent" object is thus internally protected from the fear, rage, or grief associated with the "bad parent" object. These coexisting part-selves are in turn buried under a superordinate personality structure that continues to develop. Because the part-selves remain separated, the succeeding personality structure has the potential for fragmentation, such as that occurring in multiple-personality disorders or other dissociative phenomena.

The clinical literature on addiction contains many case studies of ego states associated with part-selves. Wurmser (1985) presents an in-depth analysis of a single case that illustrates the views just presented. A drug addict whose alcoholic mother was perceived "both as a friend in childhood and as a vicious, nasty, alcoholic tormentor who had no ability to see him as an individual" (1985, 89) developed part-identities corresponding to his "two" mothers. Drug-taking was associated with a "defiant, arrogant, angry, even murderously furious man and addict," and the sober state with a "good boy, who is bending over backwards, giving in, compliant, 'well-adjusted'" (91). Especially important is the fact that extreme fluctuations in self-esteem characterized shifts from drugtaking to sobriety . Fear of rejection and consequent shame precipitated drug use. Contrition and remorse "in the form of massive shame and

guilt" brought about temporary sobriety, during which "reparation, expiation, and grandiose fantasies undoing the perceived flaws" (93) were the order of the day.

This brief vignette illustrates the central conception of this chapter: self-esteem is indeed involved in addictive substance use, but in relation to self or identity structures. Case studies do not prove theories. Yet there seems to be a preponderance of argument and evidence to suggest that positing a simple relationship between self-esteem and addictive drinking or drug use, such as the hypothesis that addicts and alcoholics are people with low self-esteem, is insufficient. Three generalizations seem in order:

1. Although it is influenced by one's current situation, self-esteem usually reflects the cohesiveness and strength of self or identity structures developed during the formative years of childhood and adolescence. Dramatic changes in self-esteem observed with substance use therefore reflect actual transformations in self or identity associated with intoxication.

2. It is further apparent that low self-esteem is not susceptible to a "quick fix," as is often assumed in prevention programs for young people. Rather, lasting enhancement of self-esteem requires the development of a positive and rigorous self-concept or identity. In the case of alcoholics and other addicts, this process requires significant changes in personality organization and associated systems of values.

3. Working directly to change levels of self-esteem in the treatment of alcoholics and addicts is not likely to be productive. Rather, the changes in self-organization necessary for attaining the "firm and coherent self" described by Kohut demand a commitment of time and effort similar to that required in the original developmental process.

Research on Self-Esteem and Substance Use

A great deal of research has been conducted concerning self-esteem and similar concepts in relation to drug and alcohol use. For the past two decades, low self-esteem has been the most popular psychological explanation of drug and alcohol abuse and addiction (Furnham and Lowick 1984). This is apparent not only in the sheer volume of work on the topic but also in the fact that various forms of self-esteem development are so frequently incorporated into alcohol and drug prevention programs for youth.

It would be neither feasible nor particularly useful to cite all of the large body of empirical research on self-esteem and alcohol and drug use produced since the 1940s. Instead, conclusions will be drawn from representative examples of this work.

Our summary of empirical research begins with a brief description of how self-esteem is actually assessed in research studies. Empirical studies of three types will then be reviewed: (1) nonexperimental research, which compares alcoholics or addicts to so-called normals in measures of self-esteem or closely related characteristics; (2) explanatory research, which asks whether alcohol and drug use enhances or maintains self-esteem under various kinds of hostile environmental conditions; and (3) research on consciousness, which assesses changes in how people feel about themselves and their environment when they are "high."

Measuring Self-Esteem

The ways in which self-esteem has been measured parallel the different ways in which the concept is defined. A common approach is to ask respondents, through multiple-choice questions, whether or not they are good at various kinds of activities. This approach assumes that people who feel competent in a sufficient number of areas will as a result experience high levels of self-esteem. That is, high self-esteem is a direct function of self-competence. Both research and theory suggest, however, that perceptions of self-competence and self-esteem are not necessarily the same thing. Bandura (1984) has researched self-competence (although he prefers the term self-efficacy ) and insists that self-competence and self-esteem are different concepts.

A second approach to measuring self-esteem, also using multiplechoice questions, is to ask respondents whether or not they like themselves in various ways. Sometimes, these questions are merely self-competence questions rephrased as "what I like about myself," but others, such as estimates of personal attractiveness, are not so closely tied to competence. This approach sees the self-concept as made up of many specific facets that, in adolescents, are organized hierarchically (Marsh and Shavelson 1985). In short, self-esteem as assessed by multiple-choice tests usually combines self-ratings of competency (efficacy) and self-acceptance or liking, in the belief that self-esteem is the sum of many specific self-assessments.

A third, alternative approach to measurement is to give the respondent a method of describing the self in a holistic way. Such an approach

attempts to determine who someone is, rather than how good he or she is, by eliciting one or more coherent self-representations. For example, the semantic differential technique (Osgood, Suci, and Tannenbaum 1957) has respondents rate themselves on a series of bipolar adjective scales such as "strong/weak." Adjective checklists (Gough 1952) have respondents check from a list those adjectives that they feel accurately describe themselves. Finally, Q-techniques require respondents to sort a set of descriptive statements about personal characteristics according to a continuum of relevance to themselves (MacAndrew 1979; Stephenson 1953).

This third group of methods tends to mix identity, or who one is, with self-esteem, or how much one likes oneself. But these approaches can be used to get more directly at self-esteem by being administered twice, once concerning the actual self ("myself as I am") and once concerning the ideal self ("myself as I would like to be"). From a strictly operational perspective, the discrepancy score between actual self-concept and ideal self-concept more closely represents self-esteem as we experience it. That is, people feel self-esteem when there is little or no discrepancy between the experienced self and the ideal self. (Admittedly, discrepancy scores are associated with technical problems in measurement, however.)

Nonexperimental Studies

One way to determine whether there is a relationship between self-esteem and drug use is to assess groups of alcoholic or other addicted persons. Studying addicted subjects rather than moderate or "controlled" drinkers and drug users is logical, because self-esteem would presumably be lowest among addicts. This approach indirectly recognizes that the use of drugs—whether they are alcohol, psychoactive medical drugs, or illicit drugs—is widely promoted and accepted in our society and that it is implausible to suggest that the majority of the population suffers from a psychological deficit. For example, Schaeffer, Schuckit, and Morrissey (1976) found no relationship between self-esteem and marijuana use among college students, a finding that is hardly surprising, as experience with marijuana is very common among college students.

Case Studies . These studies are intensive examinations of individuals or groups and ordinarily do not incorporate comparisons with other in-

dividuals or groups. Most case studies have focused on alcoholics. Findings show that self-esteem among addicted persons is extremely low, often reaching a feeling of personal worthlessness (Ghadirian 1979; Hendin 1974; Stengel 1978). Gross and Adler (1970) found that both alcoholic inpatients and members of Alcoholics Anonymous scored lower on objective tests of self-esteem than did the groups for which norms had been developed. Kinsey (1966) reported that lack of self-worth was the predominant self-characterization of alcoholic women undergoing hospital treatment. A substantial proportion disagreed with the statement, "I'm the kind of person I want to be," and believed that the people they cared most about would have agreed with their negative personal assessment.

Low levels of self-esteem have also been observed among heroin addicts (Kurtines, Hogan, and Weiss 1975; O'Mahony and Smith 1984). Similar findings have been reported for adult and adolescent polydrug addicts in treatment (Lindblad 1977; Padina and Schuele 1983). Padina and Schuele also found that heavy substance users among female junior high and high school students had relatively low self-esteem, although their levels of self-esteem were higher than those of adolescents in treatment for drug problems.

Case studies generate rather than confirm hypotheses. It can always be argued that people who have experienced the humiliation, shame, and guilt usually associated with addiction, who have been labeled as alcoholics or addicts, or who find themselves in a chemical dependency treatment program suffer from low self-esteem because of their addiction rather than because of any preexisting condition.

Criterion Group Studies . The criterion group design attempts to improve on the case study by comparing the group of interest with another group that is as similar as possible—ideally, differing only in the characteristic under study. For example, a group of alcoholics or addicts might be contrasted with a comparison group of individuals who are of the same age, educational level, and so on, but who are "normal" as far as substance use is concerned. If the matching procedure is good—and this can be a shaky assumption—this method works fairly well when the results are negative, as would be the case if alcoholics' levels of self-esteem did not differ from those of matched normals. Unfortunately, it is not much better than the case study when differences in the hypothesized direction are found. In this case, the possibility of a flawed matching process assumes great importance, for the groups could dif-

fer in some way that is unknown to the researcher but that accounts for the observed difference in self-esteem. Worse, one can still argue that the lower self-esteem of the addicted group is the result rather than the cause of the addiction.

Criterion group studies reveal that groups of problem drinkers or alcoholics score lower than normal comparison groups do on measures of self-esteem (Buhler and Lefever 1947; Button 1956a, 1956b; Maddox and Williams 1968; Schaeffer, Schuckit, and Morrissey 1976; Mayo 1979; Brown 1980). The findings are especially strong in studies comparing alcoholic and nonalcoholic women. Beckman et al. (1980) found white alcoholic women to have lower levels of self-esteem than comparable women social drinkers and male alcoholics. McLachlan et al. (1979) reported similar findings for women undergoing hospital treatment for alcoholism. Hoar (1983) found that the discrepancy between actual self-concept and ideal self-concept was greater for both female alcoholic inpatients and outpatients than for a nonalcoholic comparison group.

Heilbrun and Schwartz (1980) found that the scores of male alcoholics on a measure of self-esteem derived from the Gough adjective checklist (a global measure of identity) were lower than those of nonalcoholic controls. In this case, the same relationship did not hold for women, probably because of problems in the experimental design. Williams (1965) reported greater discrepancies between descriptions of actual and ideal selves on the Gough adjective checklist among college fraternity members who were problem drinkers than among members whose drinking did not pose any problems.

Beckman (1980) reported that a group of white female alcoholics believed that low self-esteem was the major cause of their drinking. College women who drank heavily held the same belief in a study by Beckman and Bardsley (1981). On this same theme, Anderson (1981) had alcoholic women describe their adolescence. These women characterized their adolescent selves as "full of self-doubt" (presumably implying low self-esteem) to a significantly greater degree than did their nonalcoholic sisters. The alcoholic subjects also reported feeling competent during adolescence, however, supporting Bandura's (1984) distinction between self-competence and self-esteem.

Blatt et al. (1984) compared opiate and polydrug addicts at the treatment referral stage with control and psychiatric patient groups. They found that the levels of self-esteem of the opiate addicts were lower than those of either polydrug users or nonaddicts, even though opiate

addicts saw themselves as more competent than polydrug addicts did. Moreover, self-esteem was significantly related to use of opiates by the polydrug users, with heavier opiate users reporting lower self-esteem. Graeven and Folmer (1977) likewise found level of heroin use to be negatively related to level of self-esteem. Compared to less involved users, heavy users retrospectively reported lower self-esteem during their high school years. Finally, Blatt et al. (1984) distinguished between two types of depression, the first characterized by feelings of helplessness and loneliness and the second by a sense of personal worthlessness. The latter type was manifested by the opiate addicts.

Carroll (1980) found few differences in self-esteem between male alcoholics and addicts (but excluding opiate addicts). Similar results for women alcoholics and addicts were reported by Carroll et al. (1982). These findings are not surprising, if addiction itself is the primary causal factor in low self-esteem. In other words, the personal damage associated with addiction to any substance is more or less the same for all substances except opiates. The deeply ingrained stigma conferred on heroin users by society, plus the dangerous and often criminal lifestyle associated with heroin, may set this group apart.

In a study assessing self-esteem in relation to identity, Gossop (1976) compared self-ratings by inpatient and outpatient drug addicts with self-ratings of nonusers through the semantic differential technique (Os-good, Suci, and Tannenbaum 1957). On the evaluative dimension, which used bipolar adjective scales such as "good/bad" and "important/ unimportant," the discrepancy between descriptions of actual and ideal selves was greater for female addicts than for male addicts. Addicts regardless of gender were lower on the potency dimension ("strong/weak," and so on) than were controls. Silver (1977), who obtained similar results, suggested that the potency dimension represents a theme of personal power. In this sense, addicts felt more powerless than controls did. (The experimental design cannot rule out the possibility that this feeling may result from experiences associated with the addiction itself. But this finding also anticipates a theme that will be salient when we discuss the effects of drugs on consciousness.) Feeling powerless may be a predisposing condition for low self-esteem. Coopersmith (1967) tied self-esteem directly to perceived personal power, meaning control over the events that affect one's life. Personal power is thus a perception about the self that reflects the quality of the self-concept or identity.

Back and Sullivan (1978) also used a measure of identity to assess the discrepancy between "how one appears to others" and "how one would

like to be" in a sample of middle-aged and elderly members of an insurance plan. They found that the larger the discrepancy, the more frequent was the use of medical drugs, many of which were psychoactive drugs. This relationship was stronger for women than for men.

Correlational/Analytical Studies . A third nonexperimental approach uses multiple-regression methods to assess the predictive power of self-esteem in relation to other potential predictors of substance use. This approach identifies a subset of predictors, each of which makes an independent contribution. The relative importance of each predictor is also determined.

Studies by Steffenhagen and Steffenhagen (1985) on alcoholics, Reid, Martinson, and Weaver (1987) on students in fifth through eighth grades, and Kaplan and Pokorny (1977) on high school students revealed that levels of self-esteem were consistently related inversely to substance use. That is, frequent or heavy substance users and problem users had lower self-esteem. Self-esteem was itself related to other predictor variables, however, and usually was not one of the variables selected in the multiple-regression solution. For example, Steffenhagen and Steffenhagen (1985) found that depression, rather than level of self-esteem, was the most powerful predictor of alcoholism. But low self-esteem is a major distinguishing characteristic of depression as a clinical condition, and a measure of depression may thus subsume low self-esteem, as demonstrated by Blatt et al. (1984). Kaplan and Pokorny (1977) likewise found that self-esteem was correlated with many other variables, each of which also related to substance abuse.

In general, then, the findings of correlational studies are often clouded by technical issues having to do with complex interrelationships between the measures studied.

Summary of Nonexperimental Research . Self-esteem has frequently been the focus of nonexperimental research on alcohol and drug use. There is overwhelming support for an association between low levels of self-esteem and the use of alcohol and drugs. The relationship is strong for men, especially those entering treatment programs, and strongest for women alcoholics and for opiate users regardless of sex.

Nonexperimental research designs are open to the criticism that low self-esteem in groups diagnosed as alcoholics or other drug addicts may be a result of experiences associated with the addiction rather than a preexisting or "causal" condition. The former explanation is consistent

with findings that women alcoholics and opiate addicts have the lowest self-esteem among the groups studied. Female alcoholics are likely to perceive more censure and feel more shame than will male alcoholics. Likewise, opiate addicts, usually heroin users, are the most feared and despised group of drug addicts. Nevertheless, for women alcoholics and for heroin users of both sexes there is retrospective evidence of low self-esteem during adolescence, which argues that low self-esteem may be a problem that predates addiction.

Research based on measures that assess self-esteem as the discrepancy between actual self-concept and ideal self-concept suggests that addicts also experience feelings of powerlessness. These feelings, associated with self-concept or identity, may underlie the low self-esteem characteristic of addicted people.

Explanatory Studies

A second type of research attempts to explain how drug or alcohol use "works" in relation to self-esteem. In other words, to say that people drink or use drugs because they have low self-esteem does not explain how getting drunk or high deals with the problem. Simplistic explanations—for example, suggesting that people simply "forget" their problems—are often inconsistent with the facts. An elderly recovering alcoholic told one of the authors how his anger toward other people related to getting drunk: "I drank in order to remember not to forget," he explained.

Self-Efficacy . According to Bandura (1982), the desire to be competent is a fundamental human motive. In the course of his research, self-perceptions of competence, or "self-efficacy," have been shown to influence thinking, behavior, and emotional arousal. Influential within academic psychology, Bandura's work has been the basis from which social learning theory has evolved. This theory, in turn, has of late provided a theoretical basis for prevention education, as well as for behavioral approaches to the treatment of addiction. The principle of modeling, which proposes that observing successful performance on the part of others generates an expectation of self-efficacy in the observer, has been very influential: "If they can do it, I can do it." For example, adolescents, despite threats from adults about negative consequences, probably use peer and adult models of "successful" alcohol and drug use to conclude that they will have the same capacity for control.

Bandura (1984) argues that self-efficacy and self-esteem are different. Self-efficacy refers to the evaluation of one's own competencies, whereas self-esteem refers to one's sense of self-worth. One may be a competent driver, but driving a car is something that virtually anyone can do. Or, one could be a hopelessly bad golfer, yet lack any personal investment in playing golf well. Self-efficacy would simply dictate that one should avoid playing golf.

Bandura's distinction is plausible. Still, what kinds of experiences might contribute to the development of self-esteem? It is doubtful that anyone who is not competent in activities that he or she values could feel high self-esteem. If playing golf really were important to an individual, it might be connected to self-esteem in addition to self-efficacy. Thus we are confronted with an empirical question: does self-efficacy also relate to alcohol and drug use?

Bandura (1982) readily accepts the application of self-efficacy theory to the prevention of relapse in cases of smoking, alcoholism, and other drug addictions. For example, Marlatt (1985) reported that people who felt less confident about their ability to resist resuming an addiction were in fact more likely to slip (relapse) after a period of abstinence. This finding is hardly surprising, but it may not imply what it seems to imply. Psychodynamic, as opposed to behavioral, theory could interpret such lack of confidence as resistance or rationalization, which excuses in advance the resumption of a behavior about which the individual is highly ambivalent.

Social learning theory has also been applied willy-nilly to a variety of school prevention programs in which children are led to experience success in various activities, on the theory that they will feel self-efficacious and thus not want to use drugs (Tobler 1986). A comprehensive review of such programs found no positive results as far as drug use was concerned (Schaps et al. 1983). Part of the basis for these negative results may be that the activities used in the programs may not have related to anything of personal importance to the children. Moreover, such programs are misdirected when they focus on children in general, instead of on the smaller population of children who are at high risk for addiction and other problems.

What about research that attempts to link self-efficacy to problem drinking or drug use? Vaillant (1983) found that a history of personal competence did not protect men from alcoholism. A measure of personal competence collected when his subjects were children was the best predictor of all facets of adult adjustment except later alcoholism. Even

people who were highly competent as children could become alcoholics as adults. Recall that although opiate addicts in the study by Blatt et al. (1984) had lower self-esteem than other addicts, they also described themselves as more competent than did other addict groups.

There is strong evidence that alcohol enhances feelings of personal competence. Tarbox (1979) found that male alcoholics, compared to normals, overestimated their own competence. Konovsky and Wilsnack (1982) reported that social drinkers who believed they had been drinking—both those who had actually consumed alcohol and those who had been drinking a nonalcoholic placebo—gave higher estimates of their own performance on a creativity test. Finally, anxiety about the evaluation of one's performance seems to promote more drinking. Higgins and Marlatt (1975) found that young males who were heavy social drinkers drank significantly more when they believed their performance was being evaluated by young women judges.

Perhaps these two groups of studies are complementary rather than contradictory. Vaillant's (1983) work is important because his measures were self-reports of accomplishments rather than self-ratings of competency and because the data were collected in childhood, before the onset of alcoholism. His findings support what should be obvious: highly competent people may become alcoholics or addicts. Bissell and Haberman's (1984) studies of alcoholism among professionals also document a paradox that should be readily apparent in a society in which so many talented people enter alcohol and drug treatment programs.

But being competent is not the same thing as feeling competent. The second group of studies suggests that not only alcohol but even the belief by experienced drinkers that they have consumed alcohol enhances self-estimates of competence. (By reputation, it is likely that other drugs, cocaine in particular, do the same thing.) In other words, perceived competence or self-efficacy, as contrasted to objective competence, may relate to problem drinking or drug use. Despite Bandura's plausible distinction between self-efficacy and self-esteem, it still seems likely that self-efficacy is one of the conditions that affect self-esteem, especially in relation to alcohol and drug use.

Self-Handicapping . Berglas and Jones (1978) were also interested in the relationship between alcohol use and self-efficacy, but they added the notion that people could feel competent only if they were able to attribute success to their own efforts. That is, accidental success is not a basis for self-efficacy. Success as a result of one's own efforts and compe-

tence elicits respect and love from others, which also helps to enhance self-esteem. But accidental success, or other good fortune for which one cannot take credit, is honored neither by one's self nor by others.

Based on these observations, Berglas and Jones proposed that people may use alcohol or other drugs to protect their self-image through "self-handicapping." If one is drunk, success may be perceived by others as even more remarkable, whereas failure may be excused. In effect, drinking enhances the positive attributions of others when the performance is successful but excuses failure when it is not.

In a series of experiments, Jones and Berglas (1978) allowed some of their subjects to achieve high scores on a task in a way that seemed accidental—that is, the subjects who did well were likely to attribute success to chance rather than to their own efforts. Male (but not female) subjects under this condition were more likely, on a second task, to take what they had been told was a performance-inhibiting drug. These researchers concluded that "alcohol and certain forms of drug usage may be facilitated by prior experiences of success unaccompanied by subjective feelings of mastery and control" (1978, 416). In other words, drug use could be a way of handicapping oneself in order to get more credit for success and to avoid blame for failure.

Adding the concept of attribution to self-efficacy produces a theory with more explanatory power than the concept of self-efficacy alone has. Self-handicapping in the interest of maintaining self-esteem may also be the basis for alcohol or other drug use in certain situations and by some people. It does not seem plausible, however, that this hypothesis could account for most alcohol and drug use, even by those who show signs of addiction.

Suppression of Self-Awareness . Hull proposed that alcohol consumption reduces self-awareness of performance and, in so doing, renders negative self-evaluations following failure less likely. "By decreasing the individual's level of self-awareness . . . alcohol is proposed to reduce the individual's sensitivity to potentially unfavorable information about self and hence provide a potential source of psychological relief" (1981, 594). This hypothesis is compatible with self-efficacy, in the sense that self-evaluation is tied to success or failure in performance.

Hull and Young (1983) studied alcohol consumption in relation to self-consciousness (degree of awareness and concern about one's own performance) and self-esteem in an experimental task in which subjects experienced success or failure. There were two main findings. First, the

"self-awareness suppression" hypothesis was confirmed only for highly self-conscious subjects who had been told they performed poorly. This group drank more during a wine-tasting than did either less self-conscious subjects or highly self-conscious subjects who were told they had been successful. Second, there was also an across-the-board difference related to self-esteem. Subjects with low levels of self-esteem drank more wine, regardless of either prior success or failure or high or low self-consciousness. Subjects who were highly self-conscious and had low levels of self-esteem also felt more hostility under the failure condition than subjects in any other group felt. In other words, this group not only drank more but also felt more anger.

The self-awareness hypothesis was thus only partially confirmed (for highly self-conscious subjects). In contrast, self-esteem appeared to affect drinking behavior, irrespective of other conditions. The study also revealed that self-consciousness affected drinking under conditions of failure.

Despite these interesting findings, the success-failure paradigm has serious limitations. In real life, drinking and drug use are not usually linked to a specific performance situation in which the drinker or user is being evaluated by others. Rather, they are associated with a variety of other kinds of situations and in many instances seem to occur spontaneously. The study by Hull and Young does indicate, however, that people who are both self-conscious and low in self-esteem are particularly vulnerable to failure and that individuals with low self-esteem may drink more heavily regardless of other conditions.

Self-Derogation . Kaplan (1975b) has developed and tested a theory that explicitly attributes deviant behavior in adolescents to low self-esteem. The theory assumes that adolescents who feel devalued and rejected by their peer group will develop negative attitudes about themselves. It further suggests that such socially rejected adolescents will associate the patterns of behavior endorsed by the peer group with their own negative self-evaluation: "When I try to do things their way, I always seem to fail." This association will promote deviance from group norms as a means of finding alternative ways to enhance self-esteem. Such adolescents are likely to join an "out-group" with different behavioral norms.

Kaplan and others have found considerable support for this hypothesis. Using a longitudinal research design, they found that children initially classified as "self-derogating" later adopted various types of devi-

ant behavior, including drug and alcohol use and drug dealing (Kaplan 1975a, 1976, 1977). In another study, Kaplan (1978a) found that social experiences damaging to self-esteem, in combination with associated antisocial attitudes, promoted deviant behavior.

Kaplan (1978b) also found that social class was related to the kinds of deviant behavior adopted. Self-derogating middle-class youngsters subsequently adopted behaviors considered antisocial in middle-class society. Highly self-derogating working-class students, whose group norms actually sanctioned many of those same behaviors, were likely to behave differently. Because deviance is relative to the norms of a given social group, the particular behaviors adopted in response to damaged self-esteem will vary accordingly.

Finally, Kaplan (1978a) also reported that adopting deviant behaviors worked, though it worked mainly for males. That is, among males, deviant behaviors associated with masculine roles appeared to reduce self-derogation. Among females, for whom the same behaviors were not gender-appropriate, only narcotics use by middle-class females was significantly associated with subsequent decrease in self-derogation. Among lower-class females, there was no association between deviant behavior and enhanced self-esteem.

In a study focusing specifically on the use of alcohol, Kaplan and Pokorny (1977) again found that self-enhancing effects occurred primarily for males, presumably because of role compatibility, and much more strongly where there was corresponding evidence of a personal history of rejection by peers, family, or school. These results were partly confirmed by Newcomb, Bentler, and Collins (1982) in an eight-year longitudinal study using analyses designed to establish causal relationships. These investigators found that feelings of self-derogation and alcohol use initially were positively related in the adolescents studied, yet negatively related later on, just as Kaplan and his colleagues had predicted and in part demonstrated. Despite the fact that early alcohol use decreased self-derogation later, however, adolescent self-derogation did not influence later alcohol use, possibly because the initial assessment was made at too late an age to establish this effect.

Although it is supported by evidence from longitudinal research, Kaplan's theory is at best only a partial explanation of deviant behavior in general and substance use in particular. For one thing, people who reject the norms of their social groups because they have experienced personal devaluation may opt for socially positive deviance. For example, the adolescent who is isolated or ignored in the teen social scene

may decide to be a serious student. What would account for choosing "positive," as opposed to "negative," deviance? More important, the hypothesis is significantly less applicable to the problem of substance use at a time when drug and alcohol use are no longer deviant, as is the case among teenagers and young adults today.

Despite these reservations, Kaplan's self-derogation hypothesis is relevant to at least some patterns of substance use. Frequent alcohol or drug use, multiple substance use, and the use of especially dangerous substances remain deviant, even in a population in which experimentation or occasional use of other drugs is the norm. Yet such deviant patterns of behavior are approved within subgroups that practice these behaviors. Coombs (1981) has demonstrated that heroin users follow a classic model of career achievement within the context of their own highly deviant subcultures. This kind of addiction requires the same total commitment that careers in the "straight" world require. The successful junkie who manages to obtain a regular supply of good quality heroin without engaging in "straight" behaviors such as working, who rigorously follows group norms that prohibit revealing insider information to outsiders, and who manages to elude the police achieves high status among other users and concomitant support for self-esteem.

Summary of Explanatory Research . The research reviewed does suggest a number of conditions under which alcohol or other drugs may be used to protect or enhance self-esteem. The concept of self-efficacy is important to this discussion, not only because it has been so widely applied in drug and alcohol prevention and relapse prevention, but also because its relationship to self-esteem is a matter of debate.

There is no doubt that some alcoholics and addicts are highly competent people in certain aspects of their lives and that they might feel self-efficacious as a result. But being competent is not necessarily the same as feeling competent. And being competent in one area of living, such as work, is no guarantee of competence in another area, such as sexuality. There is some evidence that alcohol enhances feelings of personal competence and that men may drink more when they are worried about the quality of their performance. In other words, it is possible that low self-efficacy may be a basis for at least some problem drinking and drug use and that self-efficacy and self-esteem overlap significantly in relation to alcohol and drug use. This impression is supported by research on self-handicapping and suppression of self-awareness. In both cases, subjects have been shown to use alcohol to protect self-esteem from the effects of

failure (low self-efficacy) in performance situations. Purely logical distinctions, like the one that can be made between self-esteem and self-efficacy, are not necessarily empirically valid distinctions.

There is also evidence that adolescents may respond to negative feedback from their social group by developing a condition of very low self-esteem, labeled "self-derogation," and that they may further associate the social norms of that group with their negative feelings about self. Joining a deviant group in which illicit drug and alcohol use is approved behavior may be a solution to this problem. Although this hypothesis about social factors underlying low self-esteem and consequent alcohol and drug use may have had considerable validity in the past, however, it seems less applicable during a period in which alcohol and drug use is relatively common in the dominant peer culture.

Research on Altered States of Consciousness

The effect of drugs, including alcohol, on self-esteem can be assessed directly by asking drinkers and users to report on their sense of self-worth both when they are sober and when they are high. Alternatively, their behavior when drinking or using drugs may be observed for signs of changes in self-esteem.

To the lay person, this kind of research might seem the most direct way to illuminate the relationship between self-esteem and substance use. People ordinarily use psychoactive drugs because drugs change the way they feel, not because they are planning to engage in some sort of task. The resulting change in consciousness may be an end in itself, or it may in addition affect the user's self-perceptions. There may be no effect on actual behavior that can be readily interpreted.

Academic psychology tends to discount or ignore research on consciousness, because it is associated with the sterile, mentalistic psychology of the late nineteenth century. Most research psychologists believe that directly observable behavioral evidence is the only evidence that counts. But this assumption is severely challenged if one assumes, as we do, that people use psychoactive drugs, including alcohol and many medical drugs, primarily to change states of consciousness.

Expectations About the Effects of Alcohol . Brown et al. (1980) developed a survey on expectations about the effects of alcohol after initial interviews with drinkers of both sexes. The survey was administered to a large sample of nonalcoholic respondents. Analysis of the results re-

vealed six groups of positive expectations, three of them (enhancement of sexual performance and experience, increased power and aggression, and increased social assertiveness) conceptually related to self-esteem. Heavier drinkers especially expected "sexual enhancement and aggressive arousal." A later study by Rohsenow (1983) of expectations by college students used a more refined method but drew similar conclusions. Heavier drinkers expected more positive consequences from drinking, as well as specific outcomes, such as social and physical pleasure, sexual enjoyment, aggressiveness, and expressiveness, that would be compatible with the enhancement of self-esteem. Unfortunately, neither of these studies assessed self-esteem directly.

Effects of Marijuana . Comprehensive studies of the drug-related experiences of marijuana users have been reported by Tart (1971) and Fisher and Steckler (1974). Neither of these studies focused specifically on self-esteem; instead, they looked at a broad spectrum of possible effects. Both studies related frequency of use to the effects experienced. Tart also related level of intoxication to drug effects.

This work, in addition to a later report by Pihl, Shea, and Costa (1979), reveals that many of the effects commonly attributed to marijuana by both experienced and inexperienced users are unrelated to self-esteem or identity. These include alterations in sensory perception, space and time perception, bodily sensations, cognitive thought processes, memory, and emotional state. Many of these effects are regularly experienced and highly valued by users. The assertion that changes in self-esteem are the primary reasons normal people use marijuana does not follow.

Nevertheless, marijuana users commonly reported positive alterations in self-esteem and identity. Tart (1971) found that a frequent effect included feeling "powerful, capable, and intelligent when stoned" and that this effect was more likely in heavier users. This observation is consistent with Silver's (1977) previously cited finding that addicts commonly felt powerless. For people who are prone to addiction, drugs may be an antidote to powerlessness and associated low self-esteem. Other marijuana-induced experiences included feeling that one's conversation is more profound and interesting, that one has deep insights into others, and that one's ideas are more original. These and other feelings are consistent with feeling powerful, but they are also consistent with what users might refer to as an "expanded" awareness, associated with a less confined or ego-centered sense of self.

In the Fisher and Steckler (1974) study, positive effects related to self-esteem and self-efficacy were invariably reported with greater frequency going up the scale from past users, occasional users, and regular users to daily users. The majority of daily users associated increased self-approval with marijuana intoxication; only about a third of regular users and one in five occasional users made this association. Similar relationships were found for other effects related to both self-esteem and self-efficacy. For example, among daily users, seven of ten felt that getting high increased their self-knowledge, and more than six of ten felt more creative and more able to communicate with others.

The fact that effects related to self-esteem are reported with increasing frequency by heavier users is most significant. Yet it will not necessarily be apparent behaviorally that an individual feels heightened self-approval, self-esteem, or creativity. Unless one is willing to accept the proposition that human beings can be powerfully motivated to attain a particular state of consciousness for its own sake, these findings may be judged as falling outside the domain of science or simply as unimportant.

Especially interesting are Tart's findings on identity changes induced by marijuana: "At higher levels of intoxication . . . the sense of separateness, of being an individual, is often replaced by feelings of oneness with the world, of actions and experiences becoming archetypal, and, occasionally, of merging with people or objects" (1971, 212). Such altered states of consciousness were greatly valued by many of Tart's respondents. Although these experiences may sound bizarre, similar states are sought and apparently attained in various forms of meditation.

The perception that the self is merging with something larger may be a solution to the problem of self-esteem. In such states, the user is no longer preoccupied with self-worth or even with the idea of self, in the sense of a finite or individual identity. Although these experiences may seem illusory to an objective observer, achieving profoundly altered states of consciousness is intensely gratifying to many marijuana and hallucinogenic drug users. Such states of consciousness may be perceived as relief from the struggle to maintain a conventional sense of identity and self-esteem. These altered states may be alternative varieties of experience that transcend our culturally approved versions of self-worth, and they may be particularly appealing to users who have problems with the culturally approved concepts.

At the same time, it should be admitted that deliberately altered states of consciousness may appeal to those who wish to experience more about themselves in a purely explorative sense. Accordingly, the

need to experience an expanded sense of reality may be a normal desire in human beings.

Effects of Alcohol on Power and Assertiveness . McClelland et al. (1972) analyzed stories made up by young male subjects about pictures of people in various situations. Behavioral data on the subjects were related to themes in the stories. Participants under both drinking and sober conditions were also observed in simulated social situations. Initial findings led to a focus on needs for personal power (n Power) and social power (s Power). Participants with high n Power scores told stories that emphasized strong, forceful actions (including assaults, giving help or advice, and controlling other people) or that expressed strong concern about reputation or prestige. This work led to an important conclusion: "Men drink primarily to feel stronger. . . . Who drinks excessively? The man who has an accentuated need for personalized power and who for a variety of reasons has chosen drinking as an outlet for it rather than some other alternative" (1972, 334).

McClelland and his associates analyzed their respondents' stories for power imagery rather than for self-esteem. Are there grounds for assuming that this work has anything to do with self-esteem? The n Power score, which consistently related to drinking, was interpreted as a measure of motivation to attain personal power. For example, the "fantasy pattern" associated with low n Power was described as follows: "Explicit concern with own reputation; also mention of superior/subordinate relationships. . . . Negative anticipations, emotions about outcomes; low prestige and view of self" (1972, 118).

Men concerned with such personalized power drank more heavily. They also often engaged in activities that created a "relatively immediate, certain, and riskless subjective feeling of power." Drinking was "a second-best alternative for the man who is both highly concerned about power and highly threatened or made fearful by it" (1972, 116). The n Power motive seems related to self-esteem; achieving n Power allows men to feel worthy.

In addition to the work by Tart (1971) and Silver (1977) already cited, other research supports the findings by McClelland et al. (1972) on feelings of power. Williams (1968) reported that alcohol consumption increased the differences between young men who were problem drinkers and those who were social drinkers. Problem drinkers became relatively more aggressive, self-centered, heedless of others, and exhibitionistic with higher doses of alcohol. These kinds of behaviors presum-

ably would be released under a heightened sense of personal power and consequent self-esteem. Parker, Gilbert, and Speltz (1981) compared male alcoholics and social drinkers concerning expectations about their own assertiveness. Social drinkers expected no change when intoxicated, whereas alcoholics expected to be more assertive. (Assertiveness implies n Power, in the sense of influencing or controlling others.)

There seems to be little doubt that male problem drinkers, and probably users of at least some other drugs, frequently feel less powerful or even powerless and that drinking or using drugs counteracts these feelings. The inference that self-esteem also varies accordingly seems inescapable.

Effects of Alcohol on Self-Esteem and Identity . A number of studies of the effects of intoxication have used adjective checklists, Q-sorts, or global assessments of discrepancy between descriptions of actual and ideal self. These studies provide direct evidence on changes in self-esteem associated with drinking.

Lang, Verret, and Watt (1984) assessed self-esteem for male and female social drinkers in both "dry" and "wet" party situations. For men, self-esteem was somewhat higher under the drinking condition than under conditions of abstinence, although the difference was not statistically significant. Women who drank were significantly lower on one measure of self-esteem. The difference between the sexes was interpreted as reflecting the relative compatibility of male and female sex roles with drinking in a social context. The important point is that for social drinkers personal and situational variables (set and setting), rather than alcohol alone, may also determine the effect of alcohol on self-esteem (Zinberg 1984).

MacAndrew and Garfinkel (1962), Vanderpool (1969), Berg (1971), and MacAndrew (1979) reported that discrepancies between actual self-concept and ideal self-concept were greater for alcoholics than for normals in the sober state and that these discrepancies increased in most respects with drinking (although Berg's 1971 results were a partial exception). In other words, when assessed as a discrepancy between actual and ideal selves, self-esteem may decline for alcoholics when they are drunk, a finding that requires some explanation.

When drunk, alcoholics also saw themselves as less submissive and eager to please and correspondingly more assertive and hostile. MacAndrew (1979) concluded that drunkenness allowed the alcoholic to dispense with any concern about accountability that might be associ-

ated with the sober self, replacing it with one's "other self," an ego state capable of expressing resentment without feeling anxious. This interpretation implies that drinking allows alcoholics to bypass their conventional identity. They are bad, but they do not care. The tyranny of negative self-esteem is circumvented by a change in identity. But the two identities are farther apart than they would be in the sober state, resulting in an apparent decline in self-esteem, if it is being assessed as discrepancy between actual and ideal selves.

Effects of Cocaine and Barbiturates on Self-Esteem and Identity . Spotts and Shontz (1984a, 1984b) used Q-sort and interview methods to identify changes in consciousness associated with heavy cocaine and barbiturate use. Instead of studying large numbers of subjects, they conducted in-depth studies of a relatively few committed male users and comparable male abstainers. Their central concept was that of ego state, defined as the "properties of the field of consciousness within which the ego is located at a given time" (1984a, 120). (As noted earlier, the activation of an ego state incorporates an associated sense of self or identity, in addition to perceptions, emotions, and motives associated with that self-concept or identity.)

Spotts and Shontz reported that moderate users of cocaine described themselves in positive terms in the pre-drug state. In contrast, heavy users in the pre-drug state experienced "massive losses in self-regard," along with increased depression and despair and violent impulses. This state corresponded to participants' characterizations of their "worst" selves. For most of these men, cocaine blocked the emergence of the worst self and delivered "a state of exhilaration, elevation of mood, and a heightened sense of well-being and mastery" (1984a, 136). Even moderate use inflated the ego and produced "exceptional feelings of exhilaration and power."

These authors also reported that moderate cocaine users were insecure and troubled by self-doubts and feelings of inadequacy. Cocaine made them feel more self-assured and seemed to enhance social relationships. In contrast, heavy users were intense, achievement-oriented men intent on self-sufficiency, loners who equated dependence on others with weakness. Their sense of being a powerful person who did not need others was reinforced by cocaine.

Spotts and Shontz (1984b) found a different picture with users of barbiturate drugs. These substances, along with alcohol, are classified as depressants. In the pre-drug state, there was a welling up of hostility

and rage accompanied by depression, anxiety, and loneliness. In the drug state, users initially felt substantially increased power and self-confidence. As this state progressed, these desired feelings were replaced by negative feelings such as hostility, arrogance, and impulsiveness. Barbiturate users became their own worst selves.

The assertiveness, hostility, and rage associated with addictive barbiturate use is reminiscent of findings in the alcohol studies cited earlier. Alcohol, cocaine, and barbiturates, though pharmacologically different, have frequently been found to dramatically enhance the male users' feelings of personal power and consequent willingness to be assertive and aggressive.

Summary of Research on Altered States of Consciousness . Studies of alcohol and other drug use reveal that self-esteem is affected by substance use, but in relation to self-concept or identity. Needs for personal power and authority appear to be associated with heavier use, almost irrespective of the drug used. Because most studies have been conducted with men, this generalization may in part reflect cultural and biological, as well as intrapsychic, influences.

It further appears that among problem drinkers and drug users the changes in self-esteem associated with intoxication may reflect a deeper transformation in self or identity. That is, the organizing principles by which the actions of others are interpreted and responded to are changed. Feelings of increased competency, aggression, and power contingent on the use of drugs thus may be derived from transformations in the perceived self. An ego state connected with intoxication replaces the ego state identified with sobriety.

For men, identity transformation may release a "bad" or "worst" self, about whom the drinker or user normally (and subsequently) feels embarrassed or ashamed. This "worst" self is less concerned about accountability and inclined to be more assertive and even hostile. Increased assertiveness suggests a concomitant increase in self-esteem. When self-esteem is assessed as a discrepancy between one's view of actual and ideal selves, however, it may appear to be lower when sober respondents reflect back on their intoxicated behavior, for the ego state associated with being drunk or high is likely to be less socially acceptable than the sober ego state.

With marijuana and some other drugs, the personality transformation may take the form of an altered state of transcendental consciousness in which the self merges with others or with the environment and

in which personal identity and fixed perspectives on reality become less important. This state may in part recapitulate the early, merged state of the infant-mother relationship (Mahler, Pine, and Bergman 1975).

Finally, major or primary identity changes associated with alcohol and drug intoxication are likely to be more characteristic of problem drinkers and users, although such changes may also occur to at least some degree in social drinkers.

Self-Esteem and Recovery

The fundamental question addressed in this chapter is whether people with a weak or fragmented sense of self or identity may sustain their self-esteem through the use of alcohol or drugs and, consequently, be especially prone to "problem" or addictive use. Undoubtedly, psychological defenses such as denial—as well as other admittedly important factors, including social supports—also play a role in alcohol or drug use, although these topics cannot be covered here.

Recovering people often refer to a period in which their addiction "worked." They suggest that as long as their habit was working, there was no reason to stop drinking or using drugs. Addictions stop being functional in a number of ways, but usually there is a history of accumulating problems. Eventually, a crisis occurs, which may be precipitated by shattered relationships, serious employment or financial problems, trouble with the law, or poor health. This crisis is the moment when drinking or using drugs can no longer suppress realistic anxiety about the problems themselves or support identity structures that maintain self-esteem. There is often a precipitous fall into demoralization (Frank 1982) and despair, the "hitting bottom" described earlier. Not surprisingly, this is also the time when alcoholics and other addicts are most likely to ask for help. For example, Matefy, Kalish, and Cantor (1971) found that alcoholics seeking help had lower self-esteem than alcoholics who rejected assistance.

All of this suggests that the recovery process should facilitate developing a personal identity that can maintain self-esteem outside the supportive environments associated with virtually all professional, as well as lay, recovery programs. Our goal is not to compare or even evaluate such programs, but rather to scrutinize pertinent research for additional evidence on the relationship between self-esteem and substance use. A recent in-depth study on recovery from heroin addiction without intervention or treatment will also be examined for the same purpose.

Treatment and Recovery Programs

Recovery programs include inpatient and outpatient medical and hospital programs; individual, family, and group psychotherapy; peer-support group programs such as Alcoholics Anonymous; and therapeutic communities, which are long-term, residential drug treatment programs incorporating a variety of peer intervention strategies within a highly structured living environment. Unfortunately, sufficient research is available only on group therapy. Methadone programs, which merely replace one drug habit with another, are not relevant. Heroin itself was once used to get people to stop using opium, but no one today would define such a practice as "treatment."

Hospital Programs . Cernovsky (1983) and Hoffman and Abbott (1970) have reported positive changes in self-esteem for inpatients in chemical dependency hospital programs. Kliner, Spicer, and Barnett (1980) reported that alcoholic physicians had higher self-esteem after hospital treatment. Other studies show less consistent results. Gross and Adler (1970) and Gross (1971) found that some aspects of self-concept related to self-esteem improved after hospital treatment, whereas other aspects did not improve or even declined. Much of this discrepancy may have resulted from differences in the measures used.

Wilson, White, and Lange (1978) compared hospital programs and community residential programs for improvement in self-esteem among skid-row alcoholics of both sexes. Significant positive gains occurred in the community programs, compared to hospital programs, although differences decreased over time. Selby (1981), studying both a hospital and a vocational rehabilitation center, found positive changes in some (but not all) aspects of self-concept that are related to self-esteem.

In a study by Heather, Rollnick, and Winton (1982), both abstainers and those who maintained "controlled" or social drinking scored higher on self-esteem measures than those who relapsed after treatment. These results were not confirmed by Wald (1980), however, who admittedly used a somewhat different research design.

With respect to changes in self-representation or identity, Armstrong and Hoyt (1963) reported that the actual or current self-perception of alcoholic males did not change during hospitalization, although the ideal self did change in the direction of a more realistic and less guilt-plagued representation. This finding is consistent with the principle that changes in identity take time. The best that could be expected from a

short-term hospital treatment program, which rarely lasts more than thirty days, is the beginning of a change in who one would like to be.

In summary, there is some evidence that self-esteem increases as a result of hospital programs, although for some groups, other approaches to treatment may result in greater gains. There is also some evidence that former patients who do not return to alcoholic drinking have higher self-esteem than those who relapse.

One study (Armstrong and Hoyt 1963) identified what may have been the beginning of a positive change in self or identity with treatment. These changes occurred only in representations of the ideal self, rather than the real self, as might be expected, given the brief nature of hospital treatment programs.

Although hospital programs are now the most common method of formal treatment for substance addictions, their usefulness, from the perspective of research that concerns us here, is seriously limited, because of their short duration. We failed to find long-term, follow-up studies of patients in hospital programs.

Therapeutic Communities . Residents who remain in therapeutic communities are usually involved in the treatment program for relatively long periods, and it is therefore possible to do genuine longitudinal research with this population. Whereas most of the research on hospital programs has focused on alcoholics, most therapeutic community residents have been addicted to other drugs, including opiate drugs. Therapeutic community residents are more likely to be from a deviant or "hard-core" user population than are alcoholics in hospital programs.

Positive changes in both self-esteem and self-concept, or identity, have been found in research on residents of therapeutic communities. Preston and Viney (1984) and Page, Mitchell, and Morris (1985) reported gains in self-esteem associated with time spent in the programs. The latter study also found smaller discrepancies in descriptions of actual selves and ideal selves, again associated with time in the program. De Leon and Jainchill (1981) found that aspects of self-concept related to self-esteem improved significantly, with women showing greater gains over time than men. These results were stable during a follow-up after two years. Similar gains were reported by Wheeler, Biase, and Sullivan (1986), although gender differences were reversed.

Because the data are longitudinal in nature and reflect a long-term treatment and recovery process, research on therapeutic community residents carries much more weight than does research on hospital patients. Here, the findings unequivocally support positive changes in self-

esteem, whether measured by self-ratings or by discrepancies between actual and ideal self-concepts. Moreover, the long period of treatment associated with therapeutic communities allows for the development of identity structures capable of maintaining self-esteem without excessive need for external support.

Alcoholics Anonymous . The membership of Alcoholics Anonymous (AA) provides another research population ideal for assessing changes in self-esteem and associated conceptions of self. The membership is large (estimated at well over one million in the United States alone). There are also many members who report long-term sobriety, making it possible to detect effects that might be apparent only after considerable time. Finally, AA insists on abstinence from alcohol and other drugs, thus eliminating problems of verification with research subjects who say they are now "controlled" drinkers or users. (The same question might be raised with respect to abstinence, but the climate of Alcoholics Anonymous is such that members who relapse soon stop participating unless they wish to get back "on the program.")

Unfortunately, we found very little research on AA members. Carroll and Fuller (1969) reported that imprisoned alcoholics who had been members of AA for some time had higher self-esteem than hospital inpatients who were being treated for alcoholism. There was no difference between AA members "on the program" and non-AA alcoholic prisoners. Conner (1962) reported that AA members showed self-concepts on Gough's (1952) adjective checklist that were more positive than those of jailed alcoholics not in AA but less positive than those of nonalcoholics. The obvious problems in research design here render the results very questionable, however.

AA members have not been studied in a way, or to a degree, that sheds light on the relationship between substance use and self-esteem, despite the fact that improving self-esteem is a frequent subject in AA and other twelve-step program meetings. Moreover, members of these programs deliberately adopt the concept of "recovering alcoholic" (or recovering addict) as an identity, which they proceed to develop over a considerable period of time. Although initially this identity may be a "public self," that is, the "image of the self one tries to convey to others" (Markus and Wurf 1987), over time it is likely to become a strong, internalized self-representation as well.

Group Therapy . In general, there is evidence that self-esteem increases although the findings are restricted to alcoholics. Gains in self-

esteem have been reported by Ends and Page (1959) for client-centered therapy patients compared with a control group, and by Annis (1979) and Annis and Chan (1983) for alcoholic prison inmates. Gad-Luther and Dickman (1979) also reported gains in self-esteem for male alcoholics participating in group therapy with their spouses. Finally, Tomsovic (1976) found gains in aspects of self-concept related to self-esteem for one method of group therapy (closed encounter) but not for another (open-group eclectic).

Summary of Research on Treatment and Recovery Programs . There is some evidence from studies of recovering alcoholics and addicts that gains in self-esteem follow various forms of treatment. There is also some reason to believe that recovering persons with higher self-esteem are less likely than others to relapse. The strongest evidence comes from studies of residents in therapeutic communities, in cases where information was collected over a relatively long period. These studies report gains in self-esteem proportional to the time residents have spent in the programs. They also report a comparable decrease in the discrepancy between residents' descriptions of their actual selves and their ideal selves, suggesting parallel development of self-concept or identity.

The lack of significant research on members of Alcoholics Anonymous and other twelve-step programs is the most glaring example of the research community's general failure to grasp the potential that this population may hold for casting light on relationships between the development of a healthy identity and associated self-esteem and recovery from addiction.

The Construction of Identity

The last, and in many respects the most illuminating, research on recovery that we will discuss comes from a study of heroin addicts who recovered without treatment. Biernacki (1986) conducted in-depth interviews of 101 heroin addicts who had been addicted for at least one year (the average was between five and six years). To be included in the study, the former addicts also had to have experienced five of the ten most common withdrawal symptoms and been free from addiction for at least two years. Finally, participants were not to have been in any formal treatment program for more than three days. This group of respondents had recovered on their own, though often with the assistance of friends or family. Their ideas about addiction and recovery represent

their own interpretations of these experiences, rather than those of treatment professionals or the members of any of the twelve-step programs.

Biernacki's research focuses on the "self-concept," or "the process of making an object of one's self and the process of role taking. Ultimately, the self is acquired through interaction with significant others" (1986, 20). The implication of this view is that, by making the self an object, human beings have the capability of gaining control over their identity. That is, should we choose to do so, we have the capacity to assume roles and act them out in relation to other people. This process results in the definition of self or identity. Self is not immutable, however; rather, "in the course of the life cycle the self evolves and changes as people alter their associations and interpret and reinterpret the actions others take in relation to them" (1986, 21).

Biernacki's approach, grounded in the sociological tradition of symbolic interactionism, is nonetheless compatible with the view of the functions of self-concept presented by Markus and Wurf (1987) or the object relations theories of Kohut (1977) or Mahler, Pine, and Bergman (1975). But it is not a theory of early development. Instead, it is a perspective on the continuous development of self during adulthood. Biernacki assumes that by "objectifying" themselves, people make choices that influence the further development of their concept of self or identity. This approach obviously runs counter to the environmental determinism of much of contemporary academic psychology, as well as to the medical disease concept of addiction.

Biernacki argues that the social role of an addict incorporates an identity that may coexist with other personal identities. Like Coombs (1981), Biernacki equates the addict's "hustle," or method of maintaining and managing a habit, to a legitimate occupation in the conventional world. If the addict's "hustle" is successful, he or she attains high social status in the world of addiction. This status in turn brings a "profound influence to bear on personal estimations of self-esteem."

Biernacki's analysis in part supports and recapitulates the thesis of this chapter. Adults, consciously or unconsciously, choose and develop conceptions of self that are tested as they are responded to by other people. When these self-conceptions are successful, in the sense of being congruent with ideal selves, character is shaped and frequently reinforced through social interactions that support self-esteem. We would also add that successful identities eventually become both encompassing, in the sense of accounting for more and more of a person's behavior, and to a considerable extent self-perpetuating. A successful and

long-maintained identity, we believe, is to a significant degree impervious to the ups and downs of daily life. The aged general still walks as if he were on parade.

In most cases, Biernacki's addicts developed a resolution to stop using heroin as a result of problems that grew out of the addiction. In about one-third of the cases, there was an emotional "bottoming out" experience. Nearly all of the other subjects developed their ideas about quitting through a rational process of considering actions and consequences. This process, too, followed an accumulation of negative experiences and often a particularly significant personal event.

If there is a single conclusion that follows from the case analyses presented, it is that "to change their lives . . . addicts must fashion new identities, perspectives, and social world involvements wherein the addict identity is excluded or dramatically depreciated" (Biernacki 1986, 141). Biernacki identifies three basic patterns by which alternative identities were developed: developing emergent identities, which were either entirely new or based on "rudimentary," preexisting identities; identity reverting, or returning to previous, unspoiled identities; and extending identities, or emphasizing and extending an identity that coexisted with the addict identity but was not contaminated by it.

These processes were universal among the addicts studied. They are illustrated by one of the examples of emergent, or new, identity. A young woman addict just released from jail had nowhere to go. Her husband was still in jail for the same drug-related offense. Her family, except for a grandmother who lived in a distant small town, had become estranged during the addiction. All of her surviving friends were practicing addicts or were in jail. With her only choices being to live with her grandmother, whom she had not seen since she became an addict, or to return to the addict culture, she took the former option.

For a considerable time, she remained in her grandmother's house, not going out because she was afraid that she would "score dope." Later, she began to leave the house, but only in the company of her grandmother. She was in an "identity hiatus," completely lacking an adult identity that could replace the identity of a drug addict. Without an adult self-conception, she remained passive, much as she would have done had she visited her grandmother's home as a child.

Eventually, her husband was released from jail and joined her. Without anything else to do, the two former big-city drug addicts went to work as laborers on a farm. After a period, they began to see themselves

as living in a directionless limbo. Then a brother-in-law suggested that they go to college. Lacking any other alternative, they went along with the suggestion.

Surprisingly, being college students provided a ready-made social role. Moreover, they met other recovering students and were invited to join a campus support group. They began to identify themselves as recovering persons rather than simply as former junkies. The extent of the development of a new identity and higher self-esteem by the time of the interview is suggested by one of the young woman's statements: "Now we're going to school. Now we feel good. Now we're really going to do something. . . . We're trying, we're kind of advertising that we're making an effort to do things that are accepted, maybe even by our parents" (Biernacki 1986, 148). Eventually, new identities developed in the addicts studied by Biernacki. This was evident when they began to refer to their addictions in the past tense. Addicts often remarked that they felt strange talking about their addictions, because it "no longer was them."

Cummings (1979) devoted his presidential address to the American Psychological Association to a description of the rules and techniques he and his colleagues developed for working with addicts in San Francisco's Haight-Ashbury district. Cummings stressed the vital importance of encouraging the addict to recall some cherished but longabandoned hope or aspiration for his or her life. The addict who could do this, in Cummings's experience, had found something positive to strive for through the long recovery period, something that, in our view, serves to revive a lost innocence and forms the basis for a positive, emergent identity.

A number of years ago, one of the authors heard a middle-aged recovering alcoholic reflect on the meaning of his sobriety. "At last," he said, "I am becoming the person I was meant to be." The implications of this statement stand at the beginning of the inquiry that led, ultimately, to this chapter.

Postscript

It remains to be determined which techniques or program ingredients are most beneficial in helping alcoholics or addicts construct healthy identities and learn associated ways of behaving that together promote and maintain healthy self-esteem. It is clear that genuinely successful in-

terventions are likely to be lengthy and to concentrate on developing personality structures that maintain self-esteem rather than on selfesteem directly. Pumping up a flat tire is an inappropriate analogy for recovery from an addiction. Rather, an internalized capability to generate self-esteem has to be developed. Recovery is no magic trick. It requires hard work in the service of significant personal development.

Bibliography

Alcoholics Anonymous. 1976. Alcoholics Anonymous . 3d ed. New York: Alcoholics Anonymous World Services.

Anderson, S. C. 1981. "Alcoholic Women: Personality Traits During Adolescence." American Journal of Drug and Alcohol Abuse 8:239–247.

Annis, H. M. 1979. "Group Treatment of Incarcerated Offenders with Alcohol and Drug Problems: A Controlled Evaluation." Canadian journal of Criminology 21:3–15.

Annis, H. M., and D. Chan. 1983. "The Differential Treatment Model: Empirical Evidence from a Personality Typology of Adult Offenders." Criminal Justice and Behavior 10:159–173.

Armstrong, R. G., and D. B. Hoyt. 1963. "Personality Structure of Male Alcoholics as Reflected in the IES Test." Quarterly Journal of Studies on Alcohol 24:239–248.

Back, K. W., and D. A. Sullivan. 1978. "Self-image, Medicine, and Drug Use." Addictive Diseases 3:373–382.

Bandura, A. 1982. "Self-Efficacy Mechanism in Human Agency." American Psychologist 37:122–147.

———. 1984. "Recycling Misconceptions of Perceived Self-Efficacy." Cognitive Therapy and Research 8:231–255.

Beck, A. T. 1967. Depression: Causes and Treatment . Philadelphia: University of Pennsylvania Press.

Beckman, L. J. 1975. "Women Alcoholics: A Review of Social and Psychological Studies." Journal of Studies on Alcohol 36:797–824.

———. 1978. "Self-Esteem of Women Alcoholics." Journal of Studies on Alcohol 39:491–498.

———. 1980. "Perceived Antecedents and Effects of Alcohol Consumption in Women." Journal of Studies on Alcohol 41:518–530.

Beckman, L. J., and P. Bardsley. 1981. "The Perceived Determinants and Consequences of Alcohol Consumption Among Young Women Heavy Drinkers." International Journal of the Addictions 16:75–88.

Beckman, L. J., T. Day, P. Bardsley, and A. Z. Seeman. 1980. "The Personality Characteristics and Family Backgrounds of Women Alcoholics." International Journal of the Addictions 15:147–154.

Berg, N. L. 1971. "Effects of Alcohol Intoxication on Self-Concept." Quarterly Journal of Studies on Alcohol 32:442–443.

Berglas, S., and E. E. Jones. 1978. "Drug Choice as a Self-Handicapping Strategy in Response to Noncontingent Success." Journal of Personality and Social Psychology 36:405–417.

Biernacki, P. 1986. Pathways from Heroin Addiction: Recovery Without Treatment . Philadelphia: Temple University Press.

Bissell, L., and P. Haberman. 1984. Alcoholism in the Professions . New York: Oxford University Press.

Blatt, S. J., B. Rounsaville, S. L. Eyre, and C. Wilber. 1984. "The Psycho-

dynamics of Opiate Addiction." Journal of Nervous and Mental Disease 172:342–352.

Brown, R. A. 1980. "Personality Measure in Gamma and Delta Alcoholics." Journal of Clinical Psychology 36:345–346.

Brown, S. A., M. S. Goldman, A. Inn, and L. R. Anderson. 1980. "Expectations of Reinforcement from Alcohol: Their Domain and Relation to Drinking Patterns." Journal of Consulting and Clinical Psychology 48:419–426.

Buhler, C., and D. W. Lefever. 1947. "A Rorschach Study on the Psychological Characteristics of Alcoholics." Quarterly Journal of Studies on Alcohol 8:197–260.

Button, A. D. 1956a. "The Psychodynamics of Alcoholism: A Survey of Eighty-Seven Cases." Quarterly Journal of Studies on Alcohol 17:443–460.

———. 1956b. "A Rorschach Study of Sixty-Seven Alcoholics." Quarterly Journal of Studies on Alcohol 17:35–52.

Carroll, J. F. X. 1980. "Similarities and Differences of Personality and Psychopathology Between Alcoholics and Addicts." American Journal of Drug and Alcohol Abuse 7:219–236.

Carroll, J. F. X., T. E. Malloy, D. L. Roscioli, G. M. Pindjak, and J. S. Clifford. 1982. "Similarities and Differences in Self-Concepts of Women Alcoholics and Drug Addicts." Journal of Studies on Alcohol 43:725–738.

Carroll, J. L., and G. B. Fuller. 1969. "The Self and Ideal-Self Concept of the Alcoholic as Influenced by Length of Sobriety and/or Participation in Alcoholics Anonymous." Journal of Clinical Psychology 25:363–364.

Cernovsky, Z. 1983. "Dimensions of Self-Actualization and Post-Treatment Alcohol Use in Fully and Partially Recovered Alcoholics." Journal of Clinical Psychology 39:628–632.

Chelton, L. G., and W. G. Bonney. 1987. "Addiction, Affects, and Self-Object Theory." Psychotherapy 24:40–46.

Conner, R. G. 1962. "The Self-Concepts of Alcoholics." In Society, Culture, and Drinking Patterns, edited by D. J. Pittman and C. R. Snyder, 455–467. New York: Wiley.

Coombs, R. H. 1981. "Drug Abuse as Career." Journal of Drug Issues 11: 369–387.

Coopersmith, S. 1967. The Antecedents of Self-Esteem . San Francisco: Freeman.

Cummings, N. A. 1979. "Turning Bread into Stones: Our Modern Antimiracle." American Psychologist 34:1119–1129.

De Leon, G., and N. Jainchill. 1981. "Male and Female Drug Abusers: Social and Psychological Status Two Years After Treatment in a Therapeutic Community." American Journal of Drug and Alcohol Abuse 8:465–497.

Ends, E. J., and C. W. Page. 1959. "Group Psychotherapy and Concomitant Psychological Change." Psychological Monographs 73(10).

Fisher, G., and A. Steckler. 1974. "Psychological Effects, Personality and Behavioral Changes Attributed to Marijuana Use." International Journal of the Addictions 9:101–126.

Frank, J. 1982. "Therapeutic Components Shared by All Psychotherapies." In Psychotherapy Research and Behavior Change, Master Lecture Series,

vol. 1, edited by J. H. Harvey and M. M. Parks, 9–37. Washington, D.C.: American Psychological Association.

Furnham, A., and V. Lowick. 1984. "Lay Theories of the Causes of Alcoholism." British Journal of Medical Psychology 57:319–332.

Gad-Luther, I., and D. Dickman. 1979. "Psychosexual Therapy with Recovering Alcoholics: A Pilot Study." Journal of Sex Education Therapy 1:11–16.

Ghadirian, A. M. 1979. "Adolescent Alcoholism: Motives and Alternatives." Comprehensive Psychiatry 20:469–474.

Gossop, M. 1976. "Drug Dependence and Self-Esteem." International Journal of the Addictions 11:741–753.

Gough, H. G. 1952. The Adjective Checklist . Palo Alto, Calif.: Consulting Psychologists Press.

Graeven, D. B., and W. Folmer. 1977. "Experimental Heroin Users: An Epidemiologic and Psychosocial Approach." American Journal of Drug and Alcohol Abuse 4:365–375.

Gross, W. F. 1971. "Self-Concepts of Alcoholics Before and After Treatment." Journal of Clinical Psychology 27:539–541.

Gross, W. F., and L. O. Adler. 1970. "Aspects of Alcoholics' Self-Concepts as Measured by the Tennessee Self-Concept Scale." Psychological Reports 27:431–434.

Hawkins, J. D., D. M. Lishner, R. F. Catalano, and M. 0. Howard. 1986. "Childhood Predictors of Adolescent Substance Abuse: Toward an Empirically Grounded Theory." Journal of Children and Contemporary Society 8:11–48.

Heather, N., S. Rollnick, and M. Winton. 1982. "Psychological Change Among Inpatient Alcoholics and Its Relationship to Treatment Outcome." British Journal on Alcohol and Alcoholism 17:90–97.

Heilbrun, A. B., and H. L. Schwartz. 1980. "Self-Esteem and Self-Reinforcement in Men Alcoholics." Journal of Studies on Alcohol 41:1134–1142.

Hendin, H. 1974. "Beyond Alienation: The End of the Psychedelic Road." American Journal of Drug and Alcohol Abuse 1:11–23.

Higgins, E. T. 1983. "A Theory of Discrepant Self-Concepts." New York University. Typescript.

Higgins, R. L., and G. A. Marlatt. 1975. "Fear of Interpersonal Evaluation as a Determinant of Alcohol Consumption in Male Social Drinkers." Journal of Abnormal Psychology 84:644–651.

Hoar, C. H. 1983. "Women Alcoholics: Are They Different from Other Women?" International Journal of the Addictions 18:251–270.

Hoffman, H., and D. Abbott. 1970. "Emotional Self-Descriptions of Alcoholic Patients After Treatment." Psychological Reports 26:892.

Hull, J. G. 1981. "A Self-Awareness Model of the Causes and Effects of Alcohol Consumption." Journal of Abnormal Psychology 90:586–600.

Hull, J. G., and R. D. Young. 1983. "Self-Consciousness, Self-Esteem, and Success-Failure as Determinants of Alcohol Consumption in Male Social Drinkers." Journal of Personality and Social Psychology 44:1097–1109.

Jones, E. E., and S. Berglas. 1978. "Control of Attributions About the Self

Through Self-Handicapping Strategies: The Appeal of Alcohol and the Role of Underachievement." Personality and Social Psychology Bulletin 4:200–206.

Kaplan, H. B. 1975a. "Increase in Self-Rejection as an Antecedent of Deviant Responses." Journal of Youth and Adolescence 4:281–292.

———. 1975b. "Sequelae of Self-Derogation." Youth and Society 7:171–197.

———. 1976. "Self-Attitudes and Deviant Response." Social Forces 54: 788–801.

———. 1977. "Antecedents of Deviant Responses: Predicting from a General Theory of Deviant Behavior." Journal of Youth and Adolescence 6:89–101.

———. 1978a. "Deviant Behavior and Self-Enhancement in Adolescence." Journal of Youth and Adolescence 7:253–277.

———. 1978b. "Social Class, Self-Derogation, and Deviant Response." Social Psychiatry 13:19–28.

Kaplan, H. B., and A. D. Pokorny. 1977. "Alcohol Use and Self-Enhancement Among Adolescents: A Conditional Relationship." In Currents in Alcoholism, vol. 4, edited by F. A. Seixas, 51–75. New York: Harcourt Brace Jovanovich.

Kernberg, O. 1966. "Structural Derivatives of Object Relationships." International Journal of Psycho-Analysis 47:236–253.

Kinsey, B. A. 1966. The Female Alcoholic: A Social Psychological Study . Springfield, Ill.: Thomas.

Kliner, D. J., J. Spicer, and P. Barnett. 1980. "Treatment Outcome of Alcoholic Patients." Journal of Studies on Alcohol 41:1217–1220.

Kobasa, S. C. 1979. "Stressful Life Events, Personality, and Health: An Inquiry into Hardiness." Journal of Personality and Social Psychology 37:1–11.

Kohut, H. 1977. The Restoration of the Self . New York: International Universities Press.

Kohut, H., and E. S. Wolf. 1978. "The Disorders of the Self and Their Treatment: An Outline." International Journal of Psycho-Analysis 59:413–425.

Konovsky, M., and S. C. Wilsnack. 1982. "Social Drinking and Self-Esteem in Married Couples." Journal of Studies on Alcohol 43:319–333.

Kurtines, W., R. Hogan, and D. Weiss. 1975. "Personality Dynamics of Heroin Use." Journal of Abnormal Psychology 84:87–89.

Lang, A. R., L. D. Verret, and C. Watt. 1984. "Drinking and Creativity: Objective and Subjective Effects." Addictive Behaviors 9:395–399.

Lindblad, R. A. 1977. "Self-Concept of Middle Socioeconomic Status Addicts: A Controlled Study." International Journal of the Addictions 12:137–151.

MacAndrew, C. 1979. "A Retrospective Study of Drunkenness: Associated Changes in the Self-Depictions of a Large Sample of Male Outpatient Alcoholics." Addictive Behaviors 4:373–381.

MacAndrew, C., and H. Garfinkel. 1962. "A Consideration of Changes Attributed to Intoxication as Common-Sense Reasons for Getting Drunk." Quarterly Journal of Studies on Alcohol 23:252–266.

McClelland, D.C., W. N. Davis, R. Kalin, and E. Wanner, eds. 1972. The Drinking Man . New York: Free Press.

McLachlan, J. F. C., R. L. Walderman, D. F. Birchmore, and L. R. Marsden.

1979. "Self-Evaluation, Role Satisfaction, and Anxiety in the Woman Alcoholic." International Journal of the Addictions 14:809–832.

Maddox, G. L., and J. R. Williams. 1968. "Drinking Behavior of Negro Collegians." Quarterly Journal of Studies on Alcohol 29:117–129.

Mahler, M. S., F. Pine, and A. Bergman. 1975. The Psychological Birth of the Human Infant . New York: Basic Books.

Markus, H., and E. Wurf. 1987. "The Dynamic Self-Concept: A Social Psychological Perspective." Annual Review of Psychology 38: 299–337.

Marlatt, G. A. 1985. "Relapse Prevention: Theoretical Rationale and Overview of the Model." In Relapse Prevention, edited by G. A. Marlatt and J. R. Gordon, 3–70. New York: Guilford.

Marsh, H. W., and R. Shavelson. 1985. "Self-Concept: Its Multifaceted, Hierarchical Structure." Educational Psychologist 20:107–123.

Matefy, R. E., R. A. Kalish, and J. M. Cantor. 1971. "Self-Acceptance in Alcoholics Who Accept and Reject Help." Quarterly Journal of Studies on Alcohol 32:1088–1091.

Mayo, E. E. 1979. The Relationship Between Self-Concept Variables and Sexual Preferences Among Male Alcoholics and Male Nonalcoholics . Ann Arbor, Mich.: University Microfilms.

Newcomb, M. D., P. M. Bender, and C. Collins. 1986. "Alcohol Use and Dissatisfaction with Self and Life: A Longitudinal Analysis of Young Adults." Journal of Drug Issues 16:479–494.

O'Mahony, P., and E. Smith. 1984. "Some Personality Characteristics of Imprisoned Heroin Addicts." Drug and Alcohol Dependence 13:255–265.

Osgood, C. E., G. J. Suci, and P. H. Tannenbaum. 1957. The Measurement of Meaning . Urbana: University of Illinois Press.

Padina, R. J., and J. A. Schuele. 1983. "Psychosocial Correlates of Alcohol and Drug Use of Adolescent Students and Adolescents in Treatment." Journal of Studies on Alcohol 44:950–973.

Page, R. C., S. Mitchell, and V. Morris. 1985. "Changes in Self-Perceptions of Illicit Drug Abusers Related to Time in Treatment in Two Residential Treatment Programs." International Journal of the Addictions 20:1741–1750.

Parker, J. C., G. Gilbert, and M. L. Speltz. 1981. "Expectations Regarding the Effects of Alcohol on Assertiveness: A Comparison of Alcoholics and Social Drinkers." Addictive Behaviors 6:29–33.

Pihl, R. O., D. Shea, and L. Costa. 1979. "Dimensions of the Subjective Marijuana Experience." International Journal of the Addictions 14:63–71.

Preston, C. A., and L. L. Viney. 1984. "Self-and Ideal-Self Perception of Drug Addicts in Therapeutic Communities." International Journal of the Addictions 19:805–819.

Reid, L. D., O. B. Martinson, and L. C. Weaver. 1987. "Factors Associated with the Drug Use of Fifth Through Eighth Grade Students." Journal of Drug Education 17:149–161.

Rohsenow, D. J. 1983. "Drinking Habits and Expectancies About Alcohol Effects for Self Versus Others." Journal of Consulting and Clinical Psychology 51:752–756.

Schaeffer, G. M., M. A. Schuckit, and E. R. Morrissey. 1976. "Correlation Be-

tween Two Measures of Self-Esteem and Drug Use in a College Sample." Psychological Reports 39:915–919.

Schaps, E., J. Moskowitz, J. Malvin, and G. Schaeffer. 1983. Napa Project Summary . Lafayette, Calif.: Pacific Institute for Research and Evaluation.

Selby, R. B. 1981. "Effects on Self-Concept in Two Different Alcoholism Treatment Programs." American Journal of Drug and Alcohol Abuse 8:95–105.

Silver, A. M. 1977. "Some Personality Characteristics of Groups of Young Drug Misusers and Delinquents." British Journal of Addiction 72:143–150.

Skager, R. W., D. G. Fisher, and E. Maddahian. 1986, A Statewide Survey of Drug and Alcohol Use Among California Students in Grades Seven, Nine, and Eleven . Sacramento, Calif.: Office of the Attorney General, Crime Prevention Center.

Spotts, J. V., and F. C. Shontz. 1984a. "Drug-Induced Ego States. I: Cocaine—Phenomenology and Implications." International Journal of the Addictions 19:119–151.

———. 1984b. "The Phenomenological Structure of Drug-Induced Ego States. II: Barbiturates and Sedative-Hypnotics—Phenomenology and Implications." International Journal of the Addictions 19:295–326.

Steffenhagen, L. A., and R. A. Steffenhagen. 1985. "Self-Esteem and Primary Demographic Characteristics of Alcoholics in a Rural State." Journal of Alcohol and Drug Education 30:51–59.

Stengel, B. 1978. "Some Observations on Repressive Values in Drug Treatment." journal of Drug Issues 8:63–73.

Stephenson, W. 1953. The Study of Behavior: Q-Technique and Its Methodology . Chicago: University of Chicago Press.

Sullivan, H. S. 1953. The Interpersonal Theory of Psychiatry . New York: Norton.

Tarbox, A. R. 1979. "Self-Regulation and Sense of Competence in Men Alcoholics." Journal of Studies on Alcohol 40:860–867.

Tart, C. T. 1971. On Being Stoned . Palo Alto, Calif.: Science and Behavior Books.

Tobler, N. S. 1986. "Meta-Analysis of 143 Adolescent Drug Prevention Programs: Quantitative Outcome Results of Program Participants Compared to a Control or Comparison Group." Journal of Drug Issues 16:537–567.

Tolpin, M. 1971. "On the Beginnings of a Cohesive Self: An Application of the Concept of Transmuting Internalization to the Study of the Transitional Object and Signal Anxiety." Psychoanalytic Study of the Child 26:316–352.

Tomsovic, M. 1976. "Group Therapy and Changes in the Self-Concept of Alcoholics." Journal of Studies on Alcohol 37:53–57.

Vaillant, G. E. 1983. The Natural History of Alcoholism . Cambridge, Mass.: Harvard University Press.

Vanderpool, J. A. 1969. "Alcoholism and the Self-Concept." Quarterly Journal of Studies on Alcohol 30:59–77.

Wald, H. P. 1980. "An Examination of the Relationship Between Drinking Status and the Self-Evaluation of the Alcoholic." Drug and Alcohol Dependence 6:285–293.

Wheeler, B. L., D. V. Biasc, and A. P. Sullivan. 1986. "Changes in Self-Concept

During Therapeutic Community Treatment: A Comparison of Male and Female Drug Abusers." Journal of Drug and Alcohol Education 16:191–196.

Williams, A. F. 1965. "Self-Concepts of College Problem Drinkers, 1: A Comparison with Alcoholics." Quarterly Journal of Studies on Alcohol 26: 586–594.

———. 1968. "Psychological Needs and Social Drinking Among College Students." Quarterly Journal of Studies on Alcohol 29:355–363.

Wilson, A., J. White, and D. E. Lange. 1978. "Outcome Evaluation of a HospitalBased Alcoholism Treatment Programme." British Journal of Addiction 73:39–45.

Wurmser, L. 1985. "Denial and Split Identity: Timely Issues in the Psychoanalytic Psychotherapy of Compulsive Drug Users." Journal of Substance Abuse Treatment 2:89–96.

Zinberg, N. E. 1985. Drug, Set, and Setting: The Basis for Controlled Intoxicant Use . New Haven, Conn.: Yale University Press.