J. DRUG EDUCATION, Vol.

20(4) 319-328,1990

CAFFEINE CONSUMPTION, EXPECTANCIES

OF CAFFEINE-ENHANCED PERFORMANCE,
AND CAFFEINISM SYMPTOMS AMONG
UNIVERSITY STUDENTS

JOHN R. BRADLEY
ALLEN PETREE

Universityof Montana

ABSTRACT
The expectancies paradigm developed in alcohol research was used to study
caffeine consumption and signs of caffeinism in two groups of college stu-
dents. A survey questionnaire was developed to gather self-report data on
expectations of caffeine-enhanced performance (EP-CAFF). level of
beverage caffeine consumed daily, and DSM-111 caffeinism signs reportedly
experienced after consumption of caffeinated beverages. A positive associa-
tion between EP-CAFF and both caffeine consumption and caffeinism signs
was predicted. In both initial (n = 527) and cross-validation (n = 270) samples,
significant correlations between EP-CAFF scores and both caffeine consump-
tion and caffeinism symptoms were found. A small, but statistically sig-
nificant, relationship was also found between caffeinism symptoms and both
alcohol consumption and cigarette smoking. Of the 797 subjects in the com-
bined sample, 151 (18.9%) endorsed five or more DSM-111 caffeinism signs.
Subjects endorsing five or more signs of caffeinism scored higher on the
EP-CAFF scale, and consumed more caffeine and alcohol in both studies and
smoked more cigarettes in study one-but not in study two-than did subjects
who endorsed fewer than five caffeinism signs. Results support the utility of
extending the expectancies model of substance use motivation from alcohol to
caffeine.

Recent psychological formulations of substance use and abuse have increasingly

pointed to motivational commonalities across diverse classes of drugs. Similar
motives for the use of different drugs is evident in current analyses of biochemical

319

0 1990,Baywood Publishing Co., Inc.

doi: 10.2190/R64X-UEMW-HE3Y-UUNA
http://baywood.com
320 1 BRADLEY AND PETREE

mechanisms of drug reinforcement [l]and the expectancies held by substance

users about the intended effects of use [2]. The present report extends a motiva-
tional model developed to study alcohol use-expctancies regarding effects-to
another widely used drug, caffeine.

WHY STUDY MOTIVES FOR CAFFEINE USE?

Besides the possible heuristic gain of extending the expectancies construct from
alcohol use to caffeine consumption, interest in caffeine is motivated by clinical
pragmatism. Caffeine consumption is ubiquitous in our society. The potential
adverse effects of caffeine consumption have become the focus of increasing
attention in recent years. The American Psychiatric Association has added caf-
feinism (caffeinism intoxication) to its Diagnostic and Statistical Manual of
Mental Disorders [31. The caffeinism syndrome includes autonomic signs,
diuresis, gastrointestinal disturbances, insomnia, and symptoms mimicking
anxiety disorders.
Caffeine-induced symptoms are of clinical significance in an of themselves.
Caffeinism is also a factor which can complicate differential diagnosis of other
psychiatric disorders [4, 51, produce additional reversible symptoms in patients
with primary psychiatric disorders [6],or form part of a constellation of multiple-
substance use [7].
Another potential concern regarding caffeine consumption which could be
relevant for college students is the possible association between maternal caffeine
consumption and birth outcomes. A review of this literature by Martin found few
studies bearing on this topic and inconsistent results in those studies reported [8].
In one carefully controlled examination of neonatal correlates of prenatal
exposure to smoking, caffeine, and alcohol, however, Jacobson, Fein, Jacobson,
Schwartz, and Dowler identified a positive association between levels of maternal
caffeine consumption and shorter gestation time, poorer neuromuscular develop-
ment, and decreased reflex functioning [9].These findings, if replicated, would
bolster the importance of understanding those factors which maintain caffeine
consumption in persons of child-bearing age.

EXPECTANCIES REGARDING CAFFEINE:

A MODEL BORROWED FROM ALCOHOL RESEARCH
In the studies reported below, we investigated the role of expectancies regarding
the performance-enhancing effects of caffeine as a motivational factor. Since the
expectancies paradigm has been valuable in predicting patterns of problem
alcohol consumption (see [lo] for a review), it is potentially useful to develop a
caffeine expectancy measure which would predict the extent of caffeine consump-
tion and the Occurrence of caffeinism symptoms.
 ENHANCED PERFORMANCE EXPECTANCIES / 321

Leigh notes that defining expectancy scale content areas on a a priori basis,
rather than clustering diverse contents together following factor analysis, makes
such scales more interpretable [lo]. A prototype of this type of scale is the
“drinking function” scale. Contents of such scales define alcohol use in terms of
the expected “functions” which alcohol is intended by the drinker to serve [ll].
We adopted a similar approach for development of the expectancy scale for
caffeinecontaining beverages reported below.

EXPECTANCIES REGARDING
CAFFEINE-ENHANCED PERFORMANCE

Enhanced performance appeared to be an appropriate candidate for an a priori

caffeine use motivation scale. For many subjects, caffeine’s effects reflect its
stimulant properties. These effects include enhanced psychomotor performance,
alertness, and relief of fatigue [12,13].
There is some experimental evidence that college students expect that caffeine
will enhance their performance. Flory and Gilbert’s review concluded that
placebo effects of caffeine increased college students performance on a variety of
tasks [14]. More recently, Kirsch and Weixel used double-blind and deceptive
administration protocols to administer decaffeinated coffee [15]. They observed
an inverted-U shaped “dose”/response curve in the performance of college student
subjects who were deceived to believe that they were consuming varying levels of
caffeine and the occurrence of caffeinism symptoms. Kirsch and Weixel’s results
clearly point to a behaviorally potent expectancy of caffeine’s enhancement of
performance.
The aim of the present studies was to develop a scale reflecting beliefs about
caffeine’s enhancement of performance with internal consistency and content
homogeneity. Based on prior findings linking positive alcohol expectancies, high
alcohol consumption, and negative drinking outcomes, it was hypothesized that
the extent of endorsement of personal expectancies of performance-enhancing
properties of caffeine-containing beverages would be positively related to level of
caffeine consumption and number of signs of caffeinism reported by each subject.
In additional analyses, as an analogue to DSM-111 classification, subjects were
divided into those reporting five or more signs of caffeine intoxication (caffeinism
“syndrome present”) and those reporting fewer than five signs (caffeinism
“syndrome absent”). These groups were then compared, hypothesizing that the
syndrome-present group would have significantly higher caffeine intake and more
performanceenhancement expectancies regarding caffeine consumption than the
syndrome-absent group.
Istivan and Matarazzo have suggested considering the overall patterns of con-
sumption of caffeine, alcohol, and cigarettes in relationship to health outcomes.
Measures of the use of cigarettes and alcohol were included in our survey and used
322 I BRADLEYANDPETREE

in exploratory analyses as predictors of caffeinism signs and caffeine intake.

Results obtained in our initial study were cross-validated in a second sample.

METHOD

Study One: Subjects and Procedures

Undergraduate student volunteers from the University of Montana and the
University of Wyoming participated in the study. University of Montana students
received experimental credit for their introductory psychology course. University
of Wyoming students agreed to participate as a course experience. Of 560 ques-
tionnaires administered, thirty-three were incomplete, improperly filled out, or
revealed that the subject did not drink caffeinated beverages. Of the total sample
of remaining 527 subjects, 290 were female and 237 were male.
The scale used to measure expectancies of enhanced performance from caffeine
consumption (EP-CAFF) contained six statements reflecting the following
reasons for drinking caffeine-containing beverages:
1. Wake up in the morning;
2. Wake up or stay awake later in the day or evening;
3. Help with study or work;
4. Improve performance;
5. Get energy; and
6. Improve concentration.
These items were included along with ten other items written to reflect the use of
caffeine containing beverages as food (e.g., tastes good with food, quenches
thirst), reflecting social setting use (when visiting with friends), or for use in
suppressing appetite. The present article is confined to analysis of the EP-CAFF
scale items. This scale’s internal consistency was judged to be adequate:
Cronbach’s alpha was .71 for study one and .73for study two.
Subjects were asked to describe their caffeinated beverage intake for the month
preceding questionnaire administration.Caffeine consumption was measured with
a scale which provides a quantity-frequency index for caffeine-containing soft-
drinks, tea, and coffee adapted from the quantity-frequency measure of alcohol
use described by Jessor, Carman, and Grossman [ll].Values assigned for caffeine
content of each serving of each.beverage were those included in DSM-111 for
soft-drinks (50 milligrams), tea (75 milligrams), and coffee (150 milligrams).
These values are the high-range values suggested in DSM-111 (American
Psychiatric Association). Average daily total caffeine (TCAFF‘) intake in milli-
grams was calculated as the sum of caffeine per day from these three beverage
sources.
Caffeinism signs were presented in checklist form using twelve “diagnostic
sign” statements modified for first-person report from the DSM-111 criteria for
 ENHANCED PERFORMANCE EXPECTANCIES / 323

caffeine intoxication. Instructions asked subjects to circle the letter next to each
item “for each change” they have noticed “after drinking caffeine-containing
beverages.” The number of signs endorsed (CAFF-SX) was entered as a con-
tinuous variable in correlational and regression analyses. Subjects were divided
into groups according to whether or not they endorsed five or more signs of
caffeinism (an analogue to the DSM-111 criterion for caffeine intoxication) for
group comparisons.
Quantity of alcohol intake from all beverage sources, expressed in ounces of
absolute alcohol per occasion of use (ETOH-QT) was calculated based upon
subjects’ questionnaire responses using the method detailed in Jessor, Carman,
and Grossman. The quantity-per occasion parameter was chosen instead of fre-
quency of drinking or average daily quantity consumed because it showed the
largest correlation with CAFF-SX.
Cigarette smokers were asked to indicate their level of use of cigarettes in
increments based on portions of a pack of cigarettes per day which were converted
to cigarettes per day (CIGSDAY) for analyses.

Study Two: Subjects and Procedures

Subjects were undergraduate psychology students from the University of Mon-
tana who volunteered to participate in the study in exchange for experimental
credit. Of 303 questionnaires administered, thirty-three were incomplete,
improperly filled out, or indicated that the subject did not drink caffeinated
beverages. Of the total of 270 remaining subjects, 178 were female and
ninety-two were male. Procedures were identical to those reported for study
one.

RESULTS
No sex differences were observed in either study one or study two for number
of caffeinism symptoms (CAFF-SX), caffeine intake (T-CAFF), or cigarettes
smoked per day (CIGSDAY). Males did report drinking greater quantities of
alcohol per drinking occasion (ETOH-QT) than females in both studies: this
difference amounted to the equivalent of an average of just over one ounce of
absolute alcohol per occasion (see Table 1).
Intercorrelations for all variables in each study are presented in Table 2. As
predicted, correlations between the EP-CAFF scale scores and T-CAFF and
CAFF-SX were positive and significant in both studies. T-CAFF was significantly
correlated with CIGSDAY in study one only. CAFF-SX was correlated sig-
nificantly with ETOH-QT in both studies, although the magnitude of this relation-
ship was greater in study two. Level of consumption of caffeine was significantly
correlated with level of consumption of alcohol and cigarette smoking in both
studies.
324 I BRADLEYANDPETREE

Table 1. Group Mean Comparisons of Male vs. Female Subjects

Study 1 Study 2

Male Female Male Female

(n = 237) (n = 290) t (n = 92) (n = 178) t

CAFF-SX
M 2.37 2.38 -.09 2.95 2.86 .26
SD 2.1 1 2.03 2.54 2.34

TCAFF
M 346.52 332.24 .57 358.97 358.10 .93
SD 289.97 281.12 369.11 257.20

EP-CAFF
M 1.31 1.50 -1.40 2.05 1.72 1.46
SD 1.60 1.53 1.86 1.56

ETOH-QT
M 5.31 3.99 5.41* 5.28 4.27 2.93*
SD 3.09 2.41 2.81 2.63

CIGS/DAY
M 1.48 1.32 .36 2.44 1.56 1.05
SD 5.38 4.78 7.31 4.92
Note: CAFF-SX = DSM-111 caffeinism signs endorsed; EP-CAFF = caffeine-enhanced
performance scale score; T-CAFF = total daily average beverage caffeine consumption
(mgs); ETOH-QT = average quantity (ounces absolute alcohol) consumed per drinking
occasion; CIGSDAY = cigarettes consumed per day (includes non-smoker subjects.
* p < .05.
* p < .01.

Stepwise multiple regression analyses were performed for each study using
EP-CAFF, T-CAFF, ETOH-QT, and CIGSDAY as independent variables and
CAFF-SX as the dependent variable. Results of these analyses are shown in Table
3. For study one, EP-CAFF, T-CAFF, and CIGSDAY yielded a multiple R of
.476 F(2,524) = 76.60,p < .001. For study two, EP-CAFF, ETOH-QT, and
T-CAFF yielded a multiple R of .468 F(3,266) = 24.92, p < .001. Though
statistically significant due to the large sample size, the additional amount of
variance in CAFF-SX beyond that associated with EP-CAFF was small for the
other variables in these equations.
Group comparisons of subjects classified as caffeinism syndrome-present (n =
87 in study one; n = 64 in study two) and syndrome-absent (n = 440 in study one;
n = 206 in study two) are shown in Table 4. In both studies syndrome-present
 ENHANCED PERFORMANCE EXPECTANCIES / 325

Table 2. lntercorrelation Matrix

Variables CAFF-SX EP-CAFF T-CAFF ETOH-OT CIGS/DAY

Study One (n = 527)

CAFF-SX .465- .270m .lo1 .173*

EP-CAFF .379- .174- .176-
T-CAFF .143- .276*
ETOH-QT .053
Study Two (n = 270)

CAFF-SX .441m .234- .206* .066

EP-CAFF .291m .208- .026
T-CAFF .089 .31Om
ETOH-QT .012
Note: CAFF-SX = DSM-111 caffeinism signs endorsed; EP-CAFF = caffeine-enhanced
performance scale score; T-CAFF = total daily average beverage caffeine consumption
(mgs); ETOH-QT = average quantity (ounces absolute alcohol) consumer per drinking
occasion; CIGS/DAY = cigarettes consumed per day (includes non-smoker subjects).
p < .05.
- p < .01.
- p < .001.

Table 3. Results of Stepwise Multiple Regression Using

Number of Caffeinism Signs as the Dependent Variable

Variables Multiple R R-Squared Beta

Study One (n = 527)

EP-CAFF .465 .216 .423

T-CAFF .476 226 .lo9
Study Two (n = 270)

EP-CAFF .441 .194 .384

ETOH-QT .456 .208 .116
T-CAFF .468 .219 .112
Note: CAFF-SX = DSM-111 caffeinism signs endorsed; EP-CAFF = caffeine-
enhanced performance scale score; T-CAFF = total daily average beverage
caffeine consumption (mgs); ETOH-QT = average uantity (ounces absolute
alcohol) consumed per drinking occasion; CIGSIDAF = cigarettes consumed
per day (includes non-smoker subjects).
326 1 BRADLEYANDPETREE

Table 4. Group Mean Comparisons of Subjects with

Five Caffeinism Signs Endorsed, Absent vs. Present

Five CaffeinismSigns

study 1 study 2
Absent Present Absent Present
(n = 440) (n = 87) t (n =207) (n = s4) t
TCAFF
M 317.67 444.83 3.81- 309.10 408.59 -2.34*
SD 266.81 345.80 299.78 289.95

EP-CAFF
M 1.16 2.70 -8.98- 1.56 2.73 -5.14-
SD 1.41 1.70 1.60 1.60

ETOH-QT
M 4.47 5.17 -2.14* 4.37 5.40 -2.68
SD 2.76 2.99 2.78 2.41

CIGS/DAY
M 1.02 3.22 3.74- 1.72 2.30 -.69
SD 4.23 7.79 5.72 6.27
Note: CAFF-SX = DSM-111 caffeinism signs endorsed; EP-CAFF = caffeine-enhanced
performance scale score; T-CAFF = total daily average beverage caffeine consumption
(mgs); ETOH-QT = average quantity (ounces absolute alcohol) consumed per drinking
occasion; CIGS/DAY = cigarettes consumed per day (indudes non-smoker subjects).
* p < .05.
*.p <.01.
-p < .01

subjects endorsed more EP-CAFF items, consumed more caffeine, and drank
greater quantities of alcohol than did syndrome-absent subjects. In study one, but
not study two, syndrome-present subjects smoked more cigarettes than did
syndrome-absent subjects.

DISCUSSION
This questionnaire survey of college students revealed an association between
the use of caffeine for the “common sense” purpose of exploiting it’s central
nervous system stimulating effects (EP-CAFFmotives) and reported levels of
both caffeine intake and DSM-111caffeinism symptoms. Our findings parallel
reports in the alcohol literature that positive expectancies regarding effects predict
both increased consumption and problem outcomes [lo].
 ENHANCED PERFORMANCE EXPECTANCIES / 327

The exploratory analyses revealed a small, but statistically significant, associa-

tion between caffeinism symptoms and quantity of alcohol consumed per
occasion. The association between cigarette smoking and caffeinism was sig-
nificant in study one, and a trend in this direction was observed in the cross
validation sample. These findings are consistent with past reports of associations
between use of caffeine, alcohol, and tobacco 171. Our data provide only modest
support, however, for possible links between cigarette smoking, drinking alcohol,
and caffeinism signs.
A substantial portion (18.9%) of our combined samples reported five or more
caffeinism signs. This finding, while clearly not a clinical appraisal of caffeinism,
bolsters the importance of studying this syndrome in college samples where the
ubiquitous use of caffeine is ingrained in the culture. The correlations observed
between endorsement of performance-enhancement motives, caffeinism
symptoms, and caffeine consumption is consistent with experimental findings that
college students expect coffee to aid their psychomotor performance 1151.
Students’ expectations of enhancing their performance through caffeine consump
tion may offer some promise as a predictor of the extent of caffeine consumption
by students and their concomitant risk of caffeinism.

REFERENCES

1. R. A. Wise, The Neurobiology of Craving: Implications for the Understanding and

Treatment of Addiction. Journal ofAbnorma1 Psychology, 97, pp. 118-132,1988.
2. T. B. Baker, Models of Addiction: Introduction to the Special Issue, Journal of
Abnormal Psychology, 97, pp. 115-117,1988.
3. American Psychiatric Association, Diagnostic andSfatisticaIManua1 ofMentalDisor-
ders, 3rd edition. Washington, D.C., 1980.
4. J. A. Sours, Case Reports of Anorexia Nervosa and Caffeinism, American Journal of
Psychiatry, 140, pp. 235-236,1983.
5. D. Behar, Flashbacks and Posttraumatic Stress Syndrome in Combat Veterans, Com-
prehensive Psychiatry, 28, pp. 459-466,1987.
6. B. A. Edelstein, C. Keaton-Brasted, and M. M. Burg, Effects of Caffeine withdrawal
on Nocturnal Enuresis, Insomnia, and Behavior Restraints, Journal of Consulting and
Clinical Psychology, 52, pp. 857-862,1984.
7. J. Istivan and J. D. Matarazzo, Tobacco, Alcohol,and Caffeine Use: A Review of Their
Interrelationships, Psychological Bulletin, 95. pp. 301-326.1984.
8. J. C.Martin, An Overview: Maternal Nicotine and Caffeine Consumption and Off-
spring Outcome, Neurobehavioral Toxicology and Teratology, 4, pp. 421-427,1982.
9. S.W.Jacobson, G. G. Fein, J. L.Jacobson, P. M.Schwartz, and J. K. Dowler, Neonatal
Correlates of Prenatal Exposure to Smoking, Caffeine, and Alcohol, Infant Behavior
and Development, 7 ,pp. 253-365,1984.
10. B. C. Leigh, In Search of the Seven Dwarves: Issues of Measurement in Alcohol
Expectancy Research, Psychological Bulletin, 105, pp. 361-373.1989.
328 / BRADLEY AND PETREE

11. R. Jessor. R. S. Carman, and P. Grossman, Expectations of Need Satisfaction and

Drinking Patterns of College Students, Quarterly Journal of Studies on Alcohol, 29,
pp. 101-116.1968.
12. S . H.Snyder and P. Sklar, Behavioral and Molecular Actions of Caffeine: Focus on
Adenosine, Journal of Psychiatric Research, 18, pp. 91-106,1984.
13. D.Sawyer, H. Julia, and A. Turin, Caffeine and Human Behavior: Arousal, Anxiety,
and Performance Effects, Journal of Behavioral Medicine, 5 , pp. 415439,1982.
14. C. Flory and J. Gilbert, The Effects of Benzedrine Sulphate and Caffeine Citrate on the
Efficiency of College Students, Journal ofApplied Psychology, 27, pp. 121-133,1943.
15. I. Kirsch and L. Weixel, Double Blind versus Deceptive Administration of a Placebo.
Behavioral Neuroscience, 102, pp. 319-323,1988.

Direct reprint requests to:

John R. Bradley, Ph.D.
Department of Psychology
University of Montana
Missoula, MT 59812