Alcohol cue reactivity and ad lib drinking in young men at risk for alcoholism

Addictive Behaviors, Vol. 15, pp. 2946, Printed in the USA. All rights reserved.

ALCOHOL

1990 Copyright

0306-4603/90 $3.00 + .OO o 1990 Pergamon Press plc

CUE REACTIVITY AND AD LIB DRINKING MEN AT RISK FOR ALCOHOLISM KIMBERLY

S. WALITZER University

and KENNETH

IN YOUNG

J. SHER

of Missouri-Columbia

Abstract - Individuals at high risk for alcoholism have been hypothesized to acquire alcoholic characteristics more rapidly than those at low risk. Two studies examined reactivity to alcohol cues, a phenomenon linked to craving for alcohol in clinical alcoholics, and ad lib drinking behavior in young men at varying risk for alcoholism. In Study 1, subjects exhibited increased autonomic and subjective responses during exposure to an alcohol beverage relative to a control beverage, suggesting that alcohol cue reactivity is not merely a clinical alcoholic phenomenon but also present in more moderate drinkers. This alcohol cue reactivity, however, was unrelated to risk status. Also, high-risk subjects exhibited greater nonspecific electromyographic and skin temperature reactivity, and higher baseline salivation volume than low-risk subjects. Of special note, ad lib alcohol consumption in Study 1 was correlated with subjects’ self-report of craving during exposure to the alcoholic beverage. Study 2 attempted to replicate the baseline salivation finding but results were equivocal concerning the robustness of this effect. Also in Study 2, subjects exhibited decreased salivation volume following a placebo beverage and increased salivation volume following alcohol consumption. Studies 1 and 2 compared ad lib drinking behavior in high-risk and low-risk samples, but no group differences were found.

An extensive literature supports the observation that alcoholism has a strong familial tendency (Cotton, 1979). Adoption studies (cf. Grove & Cadoret, 1983) have suggested that this familial pattern represents genetic transmission in at least some forms of alcoholism. In addition, familial alcoholics (i.e., alcoholics with a family history of alcoholism) have been noted to be younger when alcoholic symptoms first appear (Abelson & van der Spuy, 1978; Yamane, Katoh, & Fujita, 1980), implying that a subtype of alcoholism may be characterized by a relatively early onset of alcohol abuse. To explain the early onset in familial alcoholics, several theorists have proposed that the genetic component in alcoholism may increase the speed of the development of alcoholic symptoms (e.g., Volicer, Volicer, & D’Angelo, 1983). The mechanisms precipitating this more rapid acquisition of alcoholic characteristics are unclear. Perhaps individuals at high risk for alcoholism possess something analogous to a biological “preparedness” to develop alcohol abuse and dependence, with alcoholic symptoms developing in vulnerable individuals following relatively limited experience with ethanol. Thus, high-risk individuals may have an increased sensitivity to alcohol, leading them to develop craving and other characteristics central to alcoholism during drinking experiences which leave low-risk individuals unaffected. One “alcoholic characteristic” deserving consideration is craving for alcohol. Craving (i.e., an urgent, overpowering urge to consume alcohol) is considered a core symptom of These studies are based on a master’s thesis conducted by the first author at the University of Missouri Columbia under the supervision of the second author. Support was provided by grants AA6182 and AA7231 from the National Institute on Alcohol Abuse and Alcoholism and a Research Council Grant from the Graduate School of the University of Missouri. We wish to thank Marine11 Miller, Darcy Buerkle, Siti Noormala Rahmat, Dean Colston, and Jennifer Haynes for their assistance in the execution of these studies and Thomas M. DiLorenzo for his comments on an earlier draft of this manuscript. Correspondence concerning this article should be addressed to Kenneth J. Sher, Department of Psychology, University of Missouri - Columbia, Columbia, MO 65211.

29

30

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

alcoholism and the majority of alcoholics report craving (Ludwig & Wikler, 1974). Despite its apparently common occurrence, the conditions necessary for the development and occurrence of craving are unclear. One approach to quantifying craving involves assessing reactivity to alcohol-related cues. Research indicates that alcoholics are more reactive to alcohol-related cues than nonalcoholics, presumably due to prior conditioning. Relative to nonalcoholic control groups, during alcohol cue exposure alcoholics have exhibited increased salivation (Monti et al., 1987; Pomerleau, Fertig, Baker, & Cooney, 1983), higher levels of self-reported desire to drink (Pomerleau et al., 1983), increased self-report of physical symptoms (sweating, shaking, and heart rate increase; Cooney, Gillespie, Baker, & Kaplan, 1987) and increased skin conductance level and heart rate (Kaplan et al., 1985). Alcoholics’ salivation level during exposure to alcohol-related stimuli was positively correlated with positive expectations for alcohol consumption and expected pleasure of taste (Cooney, Baker, Pomerleau, & Josephy, 1984) and during exposure they report increased feelings of guilt and decreased self-efficacy for coping with future craving (Cooney et al., 1987). Additional data suggest that the alcoholics’ skin conductance increase associated with exposure to alcohol-related cues is correlated with severity of alcohol dependence (Kaplan, Meyer, & Stroebel, 1983; Kaplan et al., 1985). These findings indicate that craving for alcohol, assessed using salivation volume and autonomic arousal during exposure to alcohol-related cues, can be assessed objectively in alcoholics, and its intensity may be a function of dependence on alcohol and presumed reinforcement expected from alcohol. Limited research has examined reactivity to alcohol cues in social drinkers. The nonalcoholic control group in Cooney et al.‘s (1987) study evidenced increased reactivity to alcohol cues relative to control cues on several self-report measures (including increased desire to drink and expectations of pleasant effects). However, because this differential reactivity is based solely on self-report measures and is relative to a nonbeverage control substance (cedar chips), it is equivocal whether these results reflect increased alcohol-cue reactivity in social drinkers or the demand characteristics implicit in the procedure. Reactivity to alcohol cues has also been examined in social drinkers at high risk and low risk for developing drinking problems. Newlin (1985) has found increased autonomic responding in individuals at high risk for alcoholism after consuming a placebo beverage. This research suggests that social drinkers and high-risk individuals may respond differentially to alcohol cues. Based on these findings, we hypothesized that reactivity to alcohol-related cues would be greater in a sample of young men with an alcoholic parent than in a sample of young men with no alcoholic first-degree relatives. A second purpose of the current study was to examine drinking behavior in subjects at high risk and low risk for alcoholism, as alcohol consumption and drinking style are intimately linked to the development of problem drinking. Although numerous studies have examined subjective, behavioral, physiological and biochemical effects of ethanol ingestion in high-risk samples (cf. Schuckit, 1987), we know of no previous research examining ad lib drinking behavior as a dependent variable in a vulnerable sample. Given the number of findings suggesting differential sensitivity to ethanol effects, it is important to determine if risk status is related to voluntary alcohol consumption. STUDY

l:METHOD

Subjects Subjects were selected from a sample of young adult males participating in a study of the tension-reducing effects of alcohol. These subjects had been screened for the absence of self-perceived alcohol and drug problems, alcohol-related arrests, history of treatment for alcohol abuse, serious physical or psychological disorders, and use of prescription medica-

Cue reactivity

31

tions. These 21- to 26-year-old men (M = 22.4 years, SD = 1.7) had been classified as either high risk (n = 17) from having an alcoholic parent, or as low risk (n = 21) from having no alcoholic first-degree relative. Risk status for alcoholism was determined by subjects’ self-report of their parental drinking problems using versions of the Short Michigan Alcoholism Screening Test (SMAST; Selzer, Vinokur, & van Rooijen, 1975) adapted to refer to the drinking patterns of the subjects’ biological mother (M-SMAST) and father (F-SMAST). Previous research (Sher & Descutner, 1986) has shown that these reports of parental drinking behavior are a reasonably reliable method to determine risk status. Subjects who scored a five or more (three is the suggested cut-off for drinking problems) on the adapted F-SMAST or M-SMAST were classified as high-risk. Subjects whose M-SMAST and F-SMAST scores were zero were classified as low-risk. These classifications were reassessed at the end of the experimental session during a structured interview focusing on drinking problems in biological relatives. Two subjects were deleted from the sample because the results of the post-experimental interview were inconsistent with the adapted SMAST classification. The mean number of drinking occasions per month of the final sample was 10.4 (SD = 4.8; range 4-20); of these, 7.8 occasions were beer occasions and the mean number of ounces of beer per drinking occasion was 51.7 (SD = 30.1; range 12-120). Despite screening for the absence of self-perceived drinking problems, the Michigan Alcoholism Screening Test (MAST; Selzer, 1971) indicated that 28% of the sample evidenced a history of alcohol problems (scoring a 5 or greater). The drinking related variables were analyzed with t tests to assess possible initial differences between risk groups; high-risk and low-risk subjects were not found to differ on any of these drinking measures. Procedure

During the phone contact, subjects were asked to state their preferred alcoholic beverage. In order to control for the attractiveness of the beverage during the cue exposure, subjects were considered eligible if they reported beer as their preferred alcoholic beverage or if they reported drinking beer at least once a week. If the subject met these requirements, the general outline of the study was explained to him, and he was asked to report the brand of beer that he drank most frequently. If he consented to participate, he was told of the following preexperimental restrictions: (a) abstain from any type of drugs and alcohol for 24 hours before the appointment; (b) eat a meal four to six hours before the appointment and then to fast until the session; (c) refrain from physical exercise for two hours before the appointment; (d) refrain from smoking 30 minutes before the appointment; (e) not to drive a vehicle to the laboratory; and (f) make no plans for the afternoon following his appointment. On arrival at the laboratory, the subject completed an informed consent form, provided age-verifying identification, and completed an affidavit attesting to his compliance with the pre-experimental restrictions. The subject then completed a short questionnaire battery consisting of: (a) Rohsenow’s (1983) “self” version of the Alcohol Expectancy Questionnaire (AEQ), (b) the Revised Alcohol Dependence Scale (RADS; Skinner & Allen, 1982), and (c) the Restrained Drinking Scale (RDS; Ruderman & McKiman, 1984). The subject was then given a short interview administered by the experimenter to assess his alcohol dependence symptoms and quantity-frequency (QF) of drinking during the past 30 days (Polich, Armor, & Braiker, 1980). On the basis of the self-reported drinking pattern, the QF was calculated by multiplying the modal number of drinks per drinking occasion with the average number of drinking occasions per week. Subjects were then divided into High and Low QF by the total median split (Mdn = 23.4 ounces of ethanol per month). The experimenter then instructed the subject on the SHP test (Strongin, Hinsie, & Peck;

32

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

cited in Peck, 1959) procedure for salivation collection. This procedure involves the insertion of three cotton dental rolls into the mouth, one sublingually and two buccally and bilaterally. The subject was seated in an easy chair and prepared for the collection of self-report and physiological data. The measures included the following: 1. Skin temperature (ST) was measured from a thermistor attached to the distal phalanx of the fourth finger of the left hand. 2. Skin conductance level (SCL), measured using Beckman surface electrodes attached to the medial phalanges of the first and third fingers of the left hand using a conductive medium of 0.15 molar NaCl (0.9%) and two parts Unibase. Skin resistance, subsequently converted into skin conductance units, was recorded using the PGR channel (a constant current device) of a Model 7Pl22 Preamplifier. 3. Heart rate (HR) was measured using Beckman miniature surface electrodes to detect electrocardiogram from bilateral sides of the chest. The laboratory computer timed the interval between successive R-waves; prior to data analysis, these heart period data were transformed to heart rate. 4. Front&s electromyogruphic activity (EMG) was measured using Beckman miniature surface electrodes with vertical placements on both the left and right frontalis muscles. Average frontalis EMG (from both sites) was calculated prior to data analysis. 5. Salivation volume was obtained using the SHP test described above. Pre-weighed, airtight pill bottles with labeled lids containing dental rolls were placed near the subject’s right hand. 6. Self-report of subjects ’ “desire to drink” was obtained using a “desire dial” near the subject’s right hand. The face of the desire dial was a push-button telephone. The experimenter explained that the audio-tape would ask the subject the question “How much do you want to drink the beverage in front of you?” and that he should respond on the “1” (no, not at all) to “9” (yes, very much) scale. Following an eight-min adaptation period, one minute of baseline physiological recordings (excluding salivation volume) was obtained. The experimenter then gave the subject an alcohol priming dose of ,115 g ethanol/kg of body weight of beer. (The brand of beer used during each session was the brand the subject reported drinking most frequently during the earlier phone contact.) Five minutes were allotted for consumption of this relatively small priming dose (approximately 2/3 to 3/4 a can of beer for most subjects) and a five-minute absorption period followed. The rationale for using this small priming dose was to increase the subject’s comfort for consuming alcohol in the laboratory environment. When the subject had finished the priming dose, he was left alone in the experimental chamber with four opaque containers before him. Under each container was either a bottle of club soda (10 oz.) or a can of beer (12 oz.), the brand of which was previously reported as that most often consumed by the subject. The remainder of the experimental session was videotaped with a small camera mounted near the ceiling of the room. At the conclusion of the five-min absorption period, a 4.5min physiological and salivation baseline was obtained, followed by a 2.5-min rest period. Then the subject was exposed for 4.5 min to each beverage, presented in a soda-beer-soda-beer order. Due to the scarcity of high-risk subjects it was not feasible to use a counterbalanced order because this added factor would have reduced cell sizes and statistical power. This order was chosen based on previous research (Monti et al., 1987) indicating that presentation of the control beverage prior to the alcoholic beverage maximizes individual differences in salivary reactivity. The tape-recorded messages instructed the subject when to insert (just prior to turning over a container) and remove (just after covering a beverage) the dental rolls. A 4.5-min inter-exposure recording

Cue reactivity

33

period (not including salivation collection) separated each beverage exposure period. The tape-recorded verbal instructions told the subject when to turn over each container to reveal the beverage, when to open the can or bottle, when to use the “desire dial,” and provided prompts to facilitate adequate psychological exposure to the beverage. At the conclusion of the fourth exposure period, the audio-tape instructions to consume the beverage that he “would like to drink the most” were given to the subject. This beverage choice (beer or club soda) and amount consumed was recorded. The subject was then sent to the restroom, and upon return completed a post-experimental questionnaire assessing his daily and weekly beer consumption including the brand of beer used during the session. The questionnaire also assessed tobacco use, and effectiveness of the experimental manipulations. The experimenter then administered a short interview reassessing drinking problems in the subject’s biological relatives. The subject was then debriefed, paid $10.00 and transportation to his home was arranged. The portion of the videotape where the subject consumed a beverage was viewed by two trained, independent raters who recorded the number of sips (the number of times the subject brought the glass to his mouth) each subject took. The variable “sip size” was computed by dividing the total amount of beverage consumed by the number of sips the subject made to consume it. This measure assesses one aspect of drinking style (i.e., whether the subject was sipping or gulping his beverage). Interrater reliability of number of sips was calculated using intraclass correlation (Shrout & Fleiss, 1979; Formula 2) and indicated complete agreement, ICC = 1.00. Data reduction The minicomputer averaged the physiological data (HR, ST, SCL, EMG) over 30-see trials. Prior to the main analyses, a second level of data reduction averaged the physiological data over the 4.5min baselines and exposure periods. For those measures, change scores were computed for each exposure period by subtracting the previous baseline recording from each exposure period. Change scores for the salivation exposure periods were computed by subtracting the original baseline salivation volume from each of the salivation exposure periods. RESULTS

AND

DISCUSSION

Cue reactivity in high-risk versus low-risk subjects The primary dependent measures of interest were salivation volume (based on the previously reported findings of the utility of salivation volume to detect alcohol cue reactivity) and subjects’ self-reported desire to drink. The additional physiological measures were gathered to determine if other autonomic and skeletal indices of arousal would prove useful in detecting alcohol cue reactivity. Salivation volume. In order to test for initial differences in salivation volume, the salivation baseline was examined using a 2 X 2, Risk Status (high, low) X Drinking Pattern (high QF, low QF) ANOVA. A significant Risk Status effect was found, F( 1, 33) = 6.67, p < .05, indicating that high-risk subjects had a higher salivation volume (M = 2.51 g, SD = 1.30) than low-risk subjects (M = 1.66 g, SD = 0.90) prior to the beginning of the exposure periods. The interpretation for the higher salivation volume in high-risk subjects is ambiguous as subjects had an alcohol primer prior to salivation volume collection. One interpretation suggests that high salivation volume is a trait characteristic of sons of alcoholics. However, it is equally possible that the primer dose of alcohol given to the subjects differentially affected high-risk subjects (i.e., this finding represents a state characteristic of high-risk

34

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

subjects after consuming alcohol). Previous research, although sparse and inconclusive, has suggested that alcohol’s acute effect on the parotid gland is to increase salivation (Martin & Pangbom, 1971; Winsor & Strongin, 1933) and then after a short time (approximately 10 mitt), to decrease salivary flow (for 60 to 90 min; Winsor & Strongin, 1933). In the present study, the first post-alcohol salivation collection began five minutes after alcohol consumption, the final collection period began 22 minutes after consumption. The salivation volume data indicate stable levels across the collections and do not appear to reflect the above cited evidence describing a decrease in salivation 10 minutes following alcohol consumption. It is unclear whether the present data are inconsistent with this pattern of salivation volume change or if all the collection periods are in the salivation suppression period (in which case high-risk subjects’ greater salivation volume may indicate an increased level of tolerance to the salivary suppression effect relative to low-risk subjects). Visual examination of the salivation volume data suggested that an initial values effect might be present (i.e., the initial baseline value predicted subsequent variability in salivation volume). The presence of an initial values effect was confirmed by the significant correlation of the within-subjects standard deviation for the five salivation volume collection periods with the initial baseline salivation volume, r = .47, p < .Ol. Consequently, the salivation data were analyzed using percent change scores from the initial baseline salivation collection period with a 2 X 2 X 2 X 2, Risk Status X Drinking Pattern X Exposure Block (first two beverages, second two beverages) X Beverage ANOVA with the last two factors repeated measures. A significant Beverage effect was found, F(l, 33) = 4.86, p < .05, indicating that subjects exhibited a larger change from initial baseline during exposure to the beer cues (M = 18%) relative to the change from baseline during the exposure to club soda cues (M = 6%; see Figure 1). All other tests were nonsignificant. These findings indicate that this sample of moderate to heavy drinkers did evidence differential salivary reactivity to alcohol. Self-reported desire to drink. Subjects’ self-reported desire to drink each beverage was examined with a 2 X 2 X 2 X 2 X 2, Risk Status X Drinking Pattern X Exposure Block X Beverage X Repetition (beginning of exposure period, end of exposure period) ANOVA, with the last three factors repeated measures. There was a significant main effect for Beverage, F( 1, 34) = 71.34, p < ,001, with subjects reporting greater desire to drink the beer compared to the club soda (see Figure 2). Second, a significant main effect was found for Repetition, F( 1, 34) = 28.57, p < ,001 with subjects reporting greater desire to drink the beverage later in the period as opposed to earlier in the period (see Figure 2). No other effects were significant. The results of this analysis suggest that subjects found the beer to be a more attractive beverage than the club soda. In addition, subjects reported greater desire for the beverages later in the period as opposed to earlier in the period, validating the effectiveness of the exposure prompts and procedures used to elicit alcohol cue reactivity. Additional physiological measures. The I-min, preprimer baselines of the physiological measures (HR, SCL, ST, and EMG) were subjected to a 2 X 2, Risk Status X Drinking Pattern multivariate analysis of variance (MANOVA); no significant pre-existing differences were found between groups. In order to determine if the priming dose of alcohol differentially affected subjects, a 2 x 2 x 2, Risk Status X Drinking Pattern X Period (preprimer baseline, postprimer baseline) MANOVA, with the last factor a repeated measure, was performed. A significant effect for Period was found, F(4, 30) = 23.96, p < .OOl, and univariate ANOVAs revealed significant Period effects on SCL, F(1, 33) = 27.07, p < .OOl and ST, F(l, 33) = 72.55, p < ,001. The means indicate that SCL increased after the priming dose (before primer A4 = 7.18 pmhos, after primer M = 8.47 pmhos), and ST decreased (before primer M = 88.26”F, after primer M = 84.08 “F). No other tests reached significance.

35

Cue reactivity

0.01

I

1

CLUB SODA

-lO%l

BEER

CLUB SODA

BEER

BEER

CLUB SODA

BEER

I

I

CLUB SODA LEGEND:

.

*.-.-.-.-.-.-.

.

HIGH-RISK LOW-RISK

Fig. 1. Percentage of salivation volume change for high-risk and low-risk subjects during the four beverage exposure periods. Each data point indicates the percentage of change for that exposure period from the initial salivation volume baseline.

Because pre-exposure baselines had been collected prior to each exposure period, reactivity was assessed using change scores computed by subtracting out the immediately preceding baseline from each exposure period. Preliminary analyses (examination of change scores using a 2 X 2 X 2 X 2, Risk Status X Drinking Pattern X Exposure Block X Beverage MANOVA, with the last two factors repeated measures) suggested that the first

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

Q6.. 7.. 6.. 5 .. 4 .. 3 .. 2.. l-

I

:

1

2

:

CLUB SODA

0.7 0.6

:

:

:

:

1 2 CLUB SODA

L-2 BEER

l

,

t I

‘\

0.5

i

‘\

i

‘\

I

‘\

.\

‘\

‘\

z

-0.11

::

CLUB SODA

I i

!

‘\

BEER

-l.O-

0.0 -.

1.0 --

!

CLUB SODA

BEER

CLUB SODA

LEGEND:

0.0

BEER

CLUB SODA

BEER

CLUB SODA

BEER

e

HIGH-RISK

*.-.-.-.-..

LOW-RISK

1 CLUB SODA

BEER

s

Sa w$ E

/

‘\

CLUB SODA

E

I

‘\

\

W

S

/

q

1 BEER

5

i

‘\

b CLUB SODA

i

‘\

0.4

o.o-

1 2 BEER

BEER

Fig. 2. Self-report of desire to drink the beverage of high-risk and low-risk subjects, reported at the beginning and end of each exposure period on a 1 (not at all) to 9 (very much) scale. Heart rate change for high-risk and low-risk subjects during the four beverage exposure periods: Each data point indicates the change in heart rate during the exposure period from the immediately preceding baseline. Skin conductance change for high-risk and low-risk subjects during the four beverage exposure periods: Each data point indicates the change in skin conductance during the exposure period from the immediately preceding baseline. Skin temperature change for high-risk and low-risk subjects during the four beverage exposure periods: Each data point indicates the change in skin temperature during the exposure period from the immediately preceding baseline. Greater change is indicated lower on the figure so that greater reactivity is indicated higher on the figure. EMG change for high-risk and low-risk subjects during the four beverage exposure periods: Each data point indicates the change in EMG during the exposure period from the immediately preceding baseline.

Cue reactivity

31

two exposure periods were experienced as different from the last two exposure periods. Examination of Figure 2 suggests that subjects exhibited increased reactivity to the first exposure. It is possible that subjects responded to the novelty of the procedures during the first exposure period and the resulting heightened reactivity masked meaningful results. As the exposure periods become familiar this initial reactivity decreased. On the basis of this reasoning, it was decided to reanalyze the data using only the last three exposure trials. The physiological reactivity measures were reanalyzed using a 2 X 2 x 2, Risk Status x Drinking Pattern x Beverage (mean beer change score, second club soda change score) MANOVA, with the last factor a repeated measure. A significant main effect for Risk Status was found, F(4, 31) = 2.72, p < .05. The univariate ANOVAs found a significant effect for EMG, F(1, 34) = 4.30, p < .05 and a nonsignificant trend for ST, F(1, 34) = 3.33, p < .lO. Figure 2 indicates that high-risk subjects exhibited greater reactivity during the exposure periods compared to the low-risk subjects. This reactivity, although of potential interest, was nonspecific (i.e., there was no evidence of heightened reactivity to alcohol cues). Instead, high-risk subjects appeared to evidence increased EMG and ST responding to all beverage cues presented, in the absence of initial baseline differences. A significant multivariate main effect for Beverage was found, F(4, 3 1) = 8.25, p < .OOl, and univariate tests were significant for HR change, F(1, 34) = 15.82, p < .OOl and SCL change, F( 1, 34) = 26.39, p < .OOl . Inspection of Figure 2 indicates increased responsivity to the beer relative to the club soda, exhibited by greater beer change scores compared to the club soda change score. 1

Drinking behavior in high-risk versus low-risk subjects

Two measures of drinking behavior were analyzed - the total amount of beverage consumed and the sip size. Four subjects who chose to drink the club soda following the exposure periods were dropped from the following analyses as well as one additional subject who consumed both club soda and beer. These five subjects were removed from the sample so that all remaining subjects were consuming the same type of beverage (i.e., beer). The total amount of beverage consumed was calculated by subtracting the amount of beverage the subject did not consume from the amount of beverage originally present. There was a probable ceiling effect for the total amount of beverage consumed; 16 of the 38 subjects consumed their entire allotment (one can of beer). It was decided that this dependent variable would be most appropriately analyzed as a dichotomous variable finishing the beverage or not finishing the beverage. Two separate chi-square tests were performed, relating subject characteristics (Risk Status and Drinking Pattern) to completing the allotted beverage. Neither chi-square produced a significant test, indicating that alcohol cue reactivity, as measured by finishing the amount of beverage allotted, did not vary as a function of Risk Status or Drinking Pattern in this sample. Sip size was analyzed using a 2 X 2, Risk Status X Drinking Pattern ANOVA. None of the resulting tests were significant, suggesting that drinking behavior, assessed by amount of beer consumed per sip, did not differ in high-risk and low-risk subjects or in high or low QF subjects. To our knowledge, this is the first reported research to examine drinking behavior in a high- and low-risk group. Although we were unable to detect differences in drinking

‘Additional ancillary analyses were performed to identify if self-reported symptoms of alcohol dependence, positive expectancies for alcohol, and a restrained drinking style predicted autonomic reactivity to beer cues, self-report of desire for beer, and the drinking behavior measures. These analyses did not yield significant tests and thus will not be detailed further.

38

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

behavior between risk groups, the present study was potentially amount of beverage available for subjects’ consumption (1 can).

limited due to the small

Relation of cue reactivity to drinking behavior The self-reported desire to drink the beer was related to both behavioral drinking measures: the total amount of beer consumed, r = .61, p < .OOl, with subjects reporting greater desire to consume the beer drinking more beer after the exposure procedures; and the sip size, r = .35, p < .05, with subjects reporting greater desire to consume the beer consuming more beer per sip. These correlations indicate that the best predictor of consumption was the subjects’ self-report of desire for the beer during the exposure period. Thus, subjects’ conscious experience of craving the beverage was related significantly to how much they drank and their drinking style at a different time in the session. If association with later ad lib consumption is viewed as the most meaningful criterion for assessing the validity of cue-reactivity measures, the self-report measure clearly appears the most valid for this population. STUDY

2

The primary aims of Study 2 were first to replicate and explore the finding that high-risk subjects exhibit higher salivation volume than low-risk subjects, and second, to further examine ad lib drinking behavior in a high-risk and low-risk sample. In order to replicate and extend the salivation volume findings of Study 1, subjects in Study 2 consumed an alcoholic beer and a placebo beer. Based on previous research (Cox & Klinger, 1983; Keane, Lisman, & Kreutzel, 1980) indicating that nonalcoholic beers are viable placebo beverages, a nonalcoholic near beer was consumed before the alcoholic beer. If increased salivation is due to the expectancy of consuming alcohol, higher salivation should be present after consumption of the nonalcoholic beer. If increased salivation is due to ingestion of ethanol, higher salivation volume should be present only after the consumption of the alcoholic beer. In order to further examine ad lib drinking behavior in a high-risk and low-risk sample, Study 2 included an ad lib drinking period with a greater amount of beer available so that subjects could drink more heavily with fewer quantity limitations. METHOD

Subjects Twenty-nine new subjects (Cohort 2) were again selected from the pool participating in a study of the tension-reducing effects of alcohol. Because of the difficulty in building an appropriate sample size due to the relative scarcity of high-risk subjects, available subjects from Study 1 (n = 12) were recontacted (Cohort 1) and participated in Study 2. Of these 41 subjects (M = 22.31 years, SD = 1.58). 19 were classified as high-risk (including four subjects from Study 1) and 22 were classified as low-risk (including eight subjects from Study 1). Subjects’ risk status was confirmed following the experimental session with a short, structured interview and all subjects were retained in the sample. The mean number of drinking occasions per month for the sample was 9.3 (SD = 4.7, range 2-25). Of these drinking occasions, 7.0 were beer occasions and the mean number of ounces of beer per occasion was 49.4 (SD = 26.7, range 12-86). On the MAST, 28% of the sample evidenced a history of alcohol problems (using a cutoff of 5). The drinking-related variables were analyzed with t tests to assess possible initial differences between risk groups. No significant tests were found between high- and low-risk subjects for the amount of ethanol consumed in

Cue reactivity

39

a month, amount of beer consumed in a month, and for the number of days the subject drank alcohol in a month. High-risk and low-risk subjects were found to differ on the MAST, t(38) = 2.45, p < .05, with high-risk subjects scoring higher (M = 4.35) than low-risk subjects (M = 2.50). Procedure Each subject was contacted and the general outline of the study was explained to him. Interested subjects were scheduled for a two-hour appointment and told of the preexperimental restrictions described in Study 1. On arrival at the laboratory, the subject read and signed an informed consent form, provided age-verifying identification, and completed an affidavit attesting to his compliance with the pre-experimental restrictions. He then completed two brief questionnaires designed to assess subjective state: (a) Mehrabian and Russell’s (1974) Mood scale, and (b) Maisto, Connors, Tucker, McCollam, and Adesso’s (1980) Sensations scale. The subject also completed the short questionnaire battery (the AEQ, RADS, and RDS) and the interview (assessing QF and alcohol dependence symptoms based on the last 30 days) used in Study 1. As in Study 1, subjects were divided into high and low QF by the total medial split (Mdn = 2 1.5 ounces of ethanol per month). After completing the questionnaires and interview, the subject was instructed on the use of the SHP test for salivation collection, as described in Study 1. He was then seated in an easy chair and the experimenter described the next portion of the session. Several minutes after the subject had been seated alone in the experimental chamber, the experimenter started the audio-taped instructions which guided the subject through the following experimental procedures: 1. Pre-beverage salivation collection period (4.5 min). The audio-taped instructions cued the subject when to insert and remove the dental rolls for the collection period. During the collection period the subject was instructed to “sit back and relax.” 2. Placebo beer administration (3 min). A .068 g/kg dose of beverage was administered (calculated as if the beverage was 5% alcohol). Goetz Near Beer was used as the placebo beverage. When the experimenter administered the beverage, the subject was requested to answer a written question concerning the “typicality” of the beverage (“How typical is this beverage, that is, how similar does it taste to a beer that you would be served in a bar?“) on a 1 (“very different”) to 4 (“very similar”) scale. An absorption period (3 min) followed the placebo beer administration. 3. Post-placebo salivation collection period (4.5 min). The audio-taped instructions cued the subject for a salivation collection period. A waiting period (5 min) followed in order to permit the subject’s salivation volume to return to normal. During this period, the experimenter readministered the Sensations scale and the Mood scale. 4. Pre-alcohol salivation collection period (4.5 min). The audio-taped instructions cued the third salivation collection period. 5. Alcoholic beer administration (3 min).*A .068 g/kg dose of ethanol (Busch beer) was administered; the amount of liquid was equivalent to the prior placebo beer dose. During beverage administration, the subject was again asked the “typicality” question concem‘It should be noted that this design confounds the order of beverage consumption with beverage consumed. Ideally, it would have been preferable to counterbalance tire administration of the placebo beverage and the alcoholic beverage. This was not possible due to the limited number of eligible subjects. Had we increased the number of factors in our design by including order or beverage consumption, we would not have adequate power to detect order effects. We decided to administer the placebo beverage prior to the alcoholic beverage to avoid the possibility of a long-term salivation volume change induced by ethanol consumption.

KIMBERLY

40

ing the taste administration.

S. WALITZER

of the beverage.

and KENNETH

An absorption

period

J. SHER

(3 min)

followed

the beer

6. Post-alcohol salivation collection period (4.5 min). The fourth collection period was cued by the audio-taped instructions. 7. Taste-rating (10 min). At this time, the experimenter placed two unidentified pitchers of beer (Busch, Michelob Dark), two beer mugs, and two clipboards on a counter directly in front of the subject. Each clipboard contained a list of taste dimensions and the “typicality” question for the corresponding beverage. The experimenter then left the room and the tape-recorded messages instructed the subject to begin to rate these two beverages on the taste-rating scales during the next 10 minutes and to drink as much of the beverages as he would like. Then, side two of Huey Lewis and The News’ album Sports (1983) was played. The entirety of the taste-rating task was videotaped by a camera above the subject. 8. Free-drinking period (IO min). The experimenter entered the room, collected the taste-rating forms, and informed the subject that there would be a 5 to lo-minute delay. Just prior to leaving the room, the experimenter told the subject that although the taste-rating task was over he was to drink as much of the remaining beer as he liked because “it would just be thrown out afterwards.” The subject was then left alone for 10 additional minutes. At the end of the free-drinking period, the subject completed the Sensations scale and Mood scale for the third time. The subject completed a post-experimental questionnaire assessing his daily and weekly consumption of Busch, Michelob Dark, and of beer in general, his tobacco use, followed by manipulation checks of his session beer consumption and gave a breath sample for a BAC reading. The experimenter administered the structured interview to reassess family history of drinking problems in the subject’s biological relatives. The subject was then debriefed, paid $10.00, and arranged transportation home. The videotapes were viewed by two independent raters who rated the total number of sips (the number of times the subject brought the glass to his mouth) during the ad lib drinking period. The variable “sip size” was computed by dividing the total amount of beverage consumed by the number of sips the subject made. Interrater reliability for number of sips was assessed using intraclass correlation (Shrout & Fleiss, 1979; Formula 2) and indicated complete agreement, ICC = 1.00. RESULTS

Manipulation

AND

DISCUSSION

check

“Typicality” ratings of the placebo and alcoholic primer doses of beer were analyzed using a 2 X 2 x 2, Risk Status X Drinking Pattern X Beverage (placebo, alcohol) ANOVA with the last factor a repeated measure. The Beverage effect was significant, F(1, 33) = 20.95, p < ,001, and the means indicated subjects reported the alcoholic beer (M = 3.27) to be more “typical” than the placebo beer (M = 2.48). The three-way interaction was also significant, F(1, 33) = 7.03, p < .0.5, and subsequently the two-way Drinking Pattern X Beverage interaction was examined at each level of Risk Status to explore the three-way interaction. The two-way interaction was nonsignificant for the high-risk subjects (F[ 1, 331 = 2.97, p > .05) and significant for the low-risk subjects (F[l, 331 = 4.17, p < .05). The means indicate that for the high-risk subjects, the placebo beverage was equally viable for both Drinking Pattern groups (high QF M = 2.14, low QF M = 2.50). For the low-risk subjects, however, the placebo was not judged similarly by both Drinking Pattern groups (high QF M = 3.10, low QF M = 2. lo), suggesting that the low-risk, heavier drinkers

Cue

reactivity

41

found the placebo beer more “typical” than did the low-risk, lighter drinkers. The alcoholic beer appeared to be judged similarly for all groups (high-risk, high QF M = 3.57; high-risk, low QF M = 3.00; low-risk, high QF M = 3.30; low-risk, low QF M = 3.30). Salivation volume, risk status, and beverage

consumption

To assess possible initial baseline differences between groups, the pre-placebo salivation volume was subjected to a 2 X 2 X 2, Risk Status X Drinking Pattern X Cohort (Cohort 1, Cohort 2) ANOVA. Although the predicted Risk Status effect did not reach significance (and thus failed to replicate the salivation volume finding from Study l), the Risk Status X Cohort interaction approached significance, F( 1, 32) = 3.29, p < 10. The means indicate that subjects who had previously participated in Study 1 exhibited the previously demonstrated pattern of salivation volume (high-risk M = 2.91 g, low-risk M = 1.28 g); however, the new subjects did not show this same pattern (high-risk M = 2.62 g, low-risk M = 3.20 g). This partial replication suggests it is unlikely that the general failure to replicate the salivation effect in Study 2 is due to methodological differences between the two studies (e.g., dose of ethanol, quantity of beverage consumed, brand of beer, subject activity level). However, it is also unlikely that this initial finding from Study 1 is spurious, but instead reflects stable differences in that particular sample.3 No other tests reached significance. The two salivation percent change scores (placebo beverage and alcoholic beverage) were analyzed using a 2 X 2 X 2 X 2, Risk Status X Drinking Pattern X Cohort X Beverage ANOVA, with the last factor a repeated measures. A significant Beverage effect was found, F( 1, 28) = 6.53, p < .05, indicating a decrease in salivation after the alcoholic beverage consumption (M = - 15%) and an increase after the placebo beverage consumption (M = 9%; see Figure 3). A significant Risk Status x Drinking Pattern interaction was also found, F( 1,28) = 4.62, p -=c.05. The means of the percentage of salivation change scores indicated that the high-risk, high QF subjects and the low-risk, low QF subjects averaged lower percentage change scores (M = - 18% and M = - lo%, respectively), the high-risk, low QF subjects averaged a near zero change score (M = -2%), and the low-risk, high QF subjects averaged a positive change score (M = 14%). No further tests were significant. Drinking behavior and risk status

The three drinking behavior measures (the adjusted amount of beer consumed, the sip size, and the final blood alcohol concentration) were analyzed. Because the total amount of beer presented was based on body weight, the adjusted amount of beer consumed was computed by dividing the amount of beer consumed by the amount of beer allotted. These measures were subjected to a 2 X 2, Risk Status X Drinking Pattern MANOVA. None of the resulting tests were significant, indicating that these drinking behavior measures did not differ by the Risk Status or Drinking Pattern of subjects. To our knowledge, Studies 1 and 2 are the first reported experiments to compare ad lib drinking behavior in a high-risk and low-risk sample. In our studies we were unable to detect ‘To explore potential explanations for the inconsistency of this salivation effect, all subjects from Study 1 were compared to the new subjects in Study 2 on a variety of measures thought to have significance in the development of alcohol problems. These measures were available from previous data collected from the study of the tension-reducing effects of alcohol. Subjects were compared on platelet monamine oxidase (MAO) activity level, the Socialization scale (from the California Personality Inventory; Gough, 1969), the MacAndrew scale (MacAndrew. 1965). the Sensation Seeking Scale (Zuckerman, Eysenck, & Eysenck, 1978). the Eysenck Personality Inventory (Eysenck & Eysenck, 1968). Tarter’s signs of “minimal brain dysfunction” (Tarter, McBride, Buonpane, & Schneider, 1977), the MAST, drug use (marijuana, hallucinogens, cocaine, amphetamines, and central nervous system depressants) and, from the current studies, quantity-frequency of alcohol consumption. Only one analysis produced a significant finding which is most likely spurious given the number of tests performed. Presumably, subjects in Study 2 differed from those in Study 1 on some key characteristics, but we were unable to characterize these differences.

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

l. . -.._ -.. )._._._._.*._.-.-.-.~

40%

PRE-

POST-

PRE-

POST-

PLACEBO

PLACEBO

ALCOHOL

ALCOHOL

-

30% -20%-lO%-O%--lO%--

-2O%--30% PLACEBO

LEGEND:

ALCOHOL

p

HIGH-RISK,

COHORT

1

o-.-.-.-.-.-o

LOW-RISK,

COHORT

1

p

HIGH-RISK,

COHORT

2

c-,-,-.-.-.-o

LOW-RISK,

COHORT

2

Fig. 3. Percentage of salivation volume change by Risk Status and Study before and after each beverage consumption. Each data point indicates the percentage of change in the post-beverage salivation volume from the pre-beverage salivation volume.

differences between these groups in either amount of beer consumed ad lib drinking period.

or sip size during the

Relation between drinking behavior, salivation volume change, and alcohol-related questionnaire measures Salivation change scores were computed from the placebo beverage and the alcoholic

Cue reactivity

43

beverage by subtracting out the pre-beverage salivation volume from the post-beverage salivation volume. These two salivation change scores (the percent change to the placebo beverage and alcohol beverage) and the two indices of drinking behavior (adjusted amount of beer consumed, sip size) were correlated with alcohol-related questionnaire measures. Restrained drinking was significantly correlated with the percentage of salivation change due to alcohol consumption, r = - .44, p < .Ol, indicating an increase in percentage of salivation change after alcohol consumption is related to a less restrained drinking style. The adjusted amount of beer consumed was correlated with the subjects’ self-reported amount of beer typically consumed on a beer drinking occasion, r = .43, p < .Ol. The two drinking behavior measures were also correlated with a variety of additional measures (previously collected from the study of the tension-reducing effects of alcohol) thought to be related to risk for alcoholism: the Socialization scale (Gough, 1969), the MacAndrew (1965) scale, the Sensation Seeking scales (Zuckerman, Eysenck, & Eysenck, 1978), the Eysenck Personality Inventory (Eysenck & Eysenck, 1968), Tarter’s signs of “minimal brain dysfunction” (Tarter, McBride, Buonpane, & Schneider, 1977), and MAO level. The MacAndrew scale correlated with the adjusted amount of beer consumed (r = .38, p < .05). Also, the Sensation Seeking subscale “Thrill and Adventure Seeking” correlated with the adjusted amount of beer consumed (r = .36, p < .05. However, these significant correlations should be viewed cautiously as they were drawn from 42 tests. GENERAL

DISCUSSION

The primary purpose of this research program was to identify differences in cue reactivity and ad lib alcohol consumption between subjects at high risk and low risk for developing drinking problems. Specifically, we predicted high-risk subjects would exhibit greater cue reactivity, assessed by physiological measures, self-report, and drinking behavior. Neither study provided evidence for differential reactivity to exposure to alcohol-related cues or alcohol consumption between high-risk and low-risk subjects on the variety of dependent measures with considerable similarity between risk groups. Although it is possible that the consumption of the small priming dose of alcohol reduced high-risk/low-risk differences, we consider this possibility unlikely as individual differences in cue reactivity were still evident. High-risk subjects were distinguished from low-risk subjects on several variables. In Study 1, high-risk subjects exhibited greater EMG increases during cue exposure from baseline relative to low-risk subjects and a similar near-significant trend for ST change. This heightened response suggests that high-risk subjects may be more reactive to certain stimuli (e.g., beverage cues) relative to low-risk subjects. In addition, high-risk subjects from Study 1 evidenced higher levels of salivation volume than low-risk subjects. In Study 2, this effect was partially replicated - the subset of subjects who had participated in Study 1 exhibited the same effect, and new subjects did not. Similarly, higher salivation rates in alcoholics have previously been reported. Kissin, Schenker, and Schenker (1959) reported that salivary sodium concentration, a positive correlate of salivation volume, was higher in alcoholics relative to controls. Newlin, Hot&kiss, Cox, Rauscher, and Li (1988) reported that alcoholics exhibited a nonsignificant trend towards greater salivation volume @ < .136) relative to nonalcoholic controls. The results of Study 1 indicated that this sample of moderate to heavy social drinkers evidenced differential autonomic reactivity to beer cues relative to control beverage cues. The degree of increased reactivity was unrelated to subjects’ risk status. It is possible that differential cue reactivity between risk groups may be restricted to samples with a history of prolonged heavy drinking and alcohol dependence (Kaplan et al., 1983). However, it is notable that increased reactivity for alcohol cues was detected in the sample as a whole

44

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

suggesting that individual differences in cue reactivity can be assessed in social drinkers. These two studies are the first we know of to examine ad lib drinking behavior in high-risk and low-risk subjects. Although we were not able to detect differences between these two groups in their ad lib drinking behavior, we believe that further studies in this area are warranted. The current lack of findings may reflect lack of relevant motivational factors in the experimental manipulation. Previous researchers (e.g., Collins, Parks, & Marlatt, 1985; Marlatt, Kosturn, & Lang, 1975) have manipulated negative emotional states and social factors which have influenced alcohol consumption in the laboratory. Conceivably, individual differences in drinking behavior might only emerge under specific environmental conditions (e.g., aversive stimulation, heavy drinking models). In addition, a social drinking setting may also promote the expression of individual differences in drinking behavior. Ad lib drinking behavior, however, was correlated with two presumed prealcoholic personality traits - the MacAndrew scale and the “Thrill and Adventure Seeking” subscale of Sensation Seeking. Thus, although ad lib drinking behavior was not related to family history of alcoholism in the present studies, it may be premature to conclude that ad lib drinking behavior is unrelated to risk for alcoholism in young social drinkers. The current studies have several methodological limitations common to many high-risk studies involving alcohol consumption. Relatively small sample sizes (due to the relative scarcity of offspring of alcoholics) reduce the power of statistical tests making small differences between groups difficult to detect. The use of subjects’ self-report of parental drinking problems to determine risk status, although shown to be a reasonably reliable method @her & Descutner, 1986), was not validated with other reports of parental drinking problems. (Cotton, 1979, suggests that the classification of risk status based solely on subjects’ self-reports of parental drinking problems is not a serious flaw as false positives are rare, and, although false negatives are more frequent, the number of true positives is small enough so that the incidence of false negatives should also be infrequent.) A third limitation is the lack of detailed family pedigree which would allow a more refined assessment of family history of alcoholism. Cloninger (1987) has suggested that familial alcoholism may be genetically heterogenous, and this heterogeneity may mask patterns of individual differences in offspring of alcoholics. For example, Cloninger describes the Type I alcoholic as exhibiting loss of control of drinking which may lead to increased susceptibility to craving alcohol following a priming dose relative to the Type II alcoholic. Individual differences in reactivity to alcohol may be determined with the aid of more detailed family pedigrees to separate Type I versus Type II family history of alcoholism. Finally, the best predictor of laboratory drinking behavior was a self-report measure self-reported desire to drink during the exposure periods. Thus, for a moderate to heavy social drinking population, self-report appears to be the most valid measure of cue reactivity. One other study, Kaplan et al. (1983) reported a similar relationship between a self-reported measure of craving and alcohol consumption: increased self-reported craving during cue exposure significantly predicted selection of an alcoholic beverage “reward” in a sample of alcoholics. We feel the present findings are important as a replication of this relationship in a different sample (social drinkers versus alcoholics) and using different measures (ad lib drinking versus selection of alcohol). After this paper was accepted for publication, a high-risk, ad lib drinking study was published (Chipperfield & Vogel-Sprott, 1988) which found a significant interaction between family history of problem drinking (positive vs. negative) and drinking model (light VS. heavy). Although that study also failed to find a main effect for family history on ad lib drinking during a taste-rating task similar to the one used in our Study 2, high-risk subjects were found to be more sensitive to the influence of the models than low-risk subjects. This

45

Cue reactivity

finding is consistent with our speculation that differences between high- and low-risk subjects on ad lib drinking might only emerge under appropriate environmental conditions.

REFERENCES Abelson, D.S., & van der Spuy, H.I.J. (1978). The age variable in alcoholism.

Journal

of Studies on Alcohol, 39,

800-808.

Chipperfield, B., & Vogel-Sprott, M. (1988). Journal of Abnormal Psychology, 97, 423-428. Collins, R.L., Parks, G.A., & Marlatt, G.A. (1985). Social determinants of alcohol consumption: The effects of social interaction and model status on the self-administration of alcohol. Journal of Consulring and Clinical Psychology,

53,

189-200.

Cooney, R.L., Baker, L.H., Pomerleau, O.F., & Josephy, B. (1984). Salivation to drinking cues in alcohol abusers: Toward the validation of a physiological measure of craving. Addictive Behaviors, 9, 91-94. Cooney, N.L., Gillespie, R.A., Baker, L.H., & Kaplan, R.F. (1987). Cognitive change after alcohol cue exposure. Journal

of Consulting

and Clinical Psychology,

55,

150-155.

Cotton, N.S. (1979). The familial incidence of alcoholism: A review. Journal of S&dies on Alcohol, 40, 89-l 16. Cox, W.M., & Klinger, E. (1983). Discriminability of regular, light, and alcoholic and nonalcoholic near beer. Journal

of Studies on Alcohol,

44, 494-498.

Cloninger, C.R. (1987). Neurogenetic adaptive mechanisms in alcoholism. Science, 236, 410-416. Eysenck, H.J., & Eysenck, S.B. (1968). Manual for the Eysenck Personality Inventory. San Diego: Educational and Industrial Testing Service. Gough, H.G. (1969). Manualfor the California Psychological Inventory. Palo Alto, CA: Consulting Psychologists Press. Grove, W.M., & Cadoret, R.J. (1983). Genetic factors in alcoholism. In B. Kissin, & H. Begleiter (Eds.), The pathogenesis of alcoholism: Biological factors (Vol. 7, pp. 31-56). New York: Plenum Press. Kaplan, R.F.. Cooney, N.L., Baker, L.H., Gillespie, R.A., Meyer, R.E., & Pomerleau, O.F. (1985). Reactivity to alcohol-related cues: Physiological and subjective responses in alcoholics and nonptoblem drinkers. Journal of Studies on Alcohol,

46, 267-272.

Kaplan, R.G., Meyer, R.E., & Stroebel, C.F. (1983). Alcohol dependence and responsivity to an ethanol stimulus as predictors of alcohol consumption. British Journal of Addicrion, 78, 259-267. Keane, T.M., Lisman, S.A., & Kreutzel, J. (1980). Alcoholic beverages and their placebos: An empirical evaluation of expectancies. Addictive Behaviors, 5, 313-328. Kissin, B., Schenker, V., & Schenket, A. (1959). The acute effects of ethyl alcohol and chlorpromazine on certain physiological functions in alcoholics. Quarterly Journal of Studies on Alcohol, 20, 480492. Lewis, H., & “The News.” (1983). Sports. [Recording]. New York: Chrysalis Records, Inc. Ludwig, A.M., & Wikle, A. (1974). “Craving” and relapse to drink. Quarterly Journal of Studies on Alcohol, 35, 108-130.

MacAndtew,

C. (1965). The differentiation

of Studies on Alcohol,

of male alcoholic outpatients by means of the MMPI. Quarrerly Journal

26, 238-246.

Maisto, S.A., Connors, G. J., Tucker, J.A., McCollam, J.B., & Adesso, V.J. (1980). Validation of the sensations scale, a measure of subjective physiological responses to alcohol. Behaviour Research and Therapy, 18, 3741. Marlatt, G.A., Kostum, C.F., & Lang, A.R. (1975). Provocation to anger and opportunity for retaliation as determinants of alcohol consumption in social drinkers. Journal of Abnormal Psychology, 84, 652659. Martin, S., & Pangbom, R.M. (1971). Human parotid secretion in response to ethyl alcohol. Journal of Dental Research,

Mehrabian, Journal

SO, 485-490.

A.,

& Russell, J.A. (1974). Distinguishing

of Consulting

and Clinical Psychology,

anger and anxiety in terms of emotional

response factors.

42, 79-83.

Monti, P.M., Bit&off, J.A., Abrams, D.B., Zwick, W.R., Nirenbetg, T.D., & Liepman, M.R. (1987). Reactivity of alcoholics and nonalcoholics to drinking cues. Journal of Abnormal Psychology, 96, 122-126. Newlin, D.B. (1985). Offspring of alcoholics have enhanced antagonistic placebo response. Journal of Studies on Alcohol,

46, 490-494.

Newlin, D.B., Hot&kiss,

F., & Li, T.-K. (1988). Autonomic and subjective reactivity paper. an aid in the detection and measurement of depression. A M A Archives of

B., Cox, W.M., Rauschet,

to alcohol stimuli with appropriate

control stimuli. Unpublished

Peck, R.E. (1959). The SHP test General Psychiatry, 1, 354. Polich, J.M., Armor, D.J., & Braiker, H.B. (1980). The course of alcoholism: Four years ajier treatment. Santa Monica, CA: Rand Corporation. Pomerleau, O.F., Fettig, J., Baker, L., & Cooney, N. (1983). Reactivity to alcohol cues in alcoholics and non-alcoholics: Implications for a stimulus control analysis of drinking. Addictive Behaviors, 8, l-10. Rohsenow, D.J. (1983). Drinking habits and expectancies about alcohol’s effects for self versus others. Journal of Consulting and Clinical Psychology, 51, 752-756. Ruderman, A.F., & McKiman, D.J. (1984). The development of a restrained drinking scale: A test of the

46

KIMBERLY

S. WALITZER

and KENNETH

J. SHER

abstinence violation effect among alcohol users. Addicrive Behaviors, 9, 264-271. Seizer, M. (1971). The Michigan alcoholism screening test: The quest for a new diagnostic Journal

of

Psychiatry,

172,

instrument.

American

1653-1658.

Seizer. M., Vinokur, A., & von Rooijen, L. (1975). A self-administered Short Michigan Alcoholism Screening Test (SMAST). America1 Journal ofPsychiatry, 172, 1653-1658. Schuckit, M.A. (1987). Biological vulnerability to alcoholism. Journal of Consulting and Clinical Psychology, 55, 301-309.

Sher,

K.J.,

& Descutner, C. (1986). Reports 11, 25-30. P.E., & Fleiss, J.L. (1979). Intraclass

of paternal

alcoholism:

Reliability

across

siblings.

Addictive

Behaviors,

Shrout,

Bulletin,

Skinner,

correlations:

Uses in assessing

rater reliability.

H.A.,

& Allen, B.A. (1982). Alcohol dependence syndrome: Measurement 91, 199-209. Tarter, R.F., McBride, H., Buonpane, N., & Schneider, D.U. (1977). Differentiation Abnormal General

Psychological

86, 420-428.

and validation.

Journal

of

of alcoholics.

Archives

of

Psychology,

Psychiatry,

34, 761-768.

Volicer, B.J., Volicer, L., & D’Angelo, N. (1983). Variation in length of time to development of alcoholism by family history of problem drinking. Drugs and Alcohol Dependence, 12, 69-83. Winsor, A.L.. & Strongin, E.I. (1933). The effect of alcohol on the rate of parotid secretion. Journal of Experimental

Psychology,

16, 589-597.

Yamane, H., Katoh, N., & Fujita, T. (1980). Characteristics of three groups of men alcoholics differentiated by age at first admission for alcoholism treatment in Japan. Journal of Studies on Alcohol, 41, 100-103. Zuckerman. M., Eysenck, S., & Eysenck, H.J. (1978). Sensation seeking in England and America: Cross-cultural, age, and sex comparisons. Journal of Consulting and Clinical Ps_whology, 46, 139-149.