is there a general tendency to become addicted?

Addicrive Behaviors, Vol. 18, pp. 8 l-87, 1993 Printed in the USA. All rights reserved.

0306-4603/93 $6.00 + .OO Copyright 0 1993 Pergamon Press Ltd.

IS THERE A GENERAL TENDENCY TO BECOME ADDICTED?

PAUL ROZIN and CARYN STOESS University of Pennsylvania

Abstract - The tendency to become addicted across a number ofdifferent substances or activities was determined for a sample of 573 subjects, including college students and their parents. Four components ofaddiction were defined: craving, tolerance, withdrawal and lack ofcontrol. Subjects rated the extent to which each of these components characterized their relationships to each often substance/activities: coffee, tea, cola beverages, favorite alcoholic beverage, chocolate, nonchocolate sweets, hot chili pepper on food, cigarettes, gambling and video games. An “addiction score” was computed for each subject and each substance/activity, by summing the scores on the four components. Correlations in addiction scores for almost all activities were positive, but low (between 0 and .30), with the exception ofchocolateand nonchocolate sweets, where the correlation was higher. The results suggest, at best, a weak tendency to become addicted, across a wide range ofsubstances or activities. Other explanations for the low positive correlations are available, besides the notion of a general tendency to become addicted. There were a few significant mother-father correlations in various addiction scores, but none between mid-parent and child values. Three of the four components of addiction (craving, lack of control and withdrawal), were highly correlated. We conclude that there is little basis for the assumption of a general tendency to become addicted. a conclusion which casts doubt on the derivative notion of an addictive personality.

The idea of a general tendency to become addicted has never, to the knowledge of the authors, been directly tested. However, investigations of the addictive personality, of which there have been many, assume that there is a tendency to become addicted that must be explained.

In general, studies of the addictive personality, which usually use addiction to one or a few of the clinically significant addictive substances (most frequently, alcohol or opiates) have not yielded a consistent result. Reviews suggest that no single personality variable is reliably associated with addiction (Cox, 1985; Kagan, 1987; Lang, 1983; Sut- ker & Allain, 1988). Antisocial tendencies, impulsivity, adventure seeking and other traits are sometimes correlated with specific addictions. One author (Nathan, 1988) concludes that almost nothing correlates reliably with alcohol or drug abuse, except antisocial behavior in childhood or adolescence, and that is not a personality variable.

There is abundant evidence of correlations between addictions to some of the major illegal or problem substances (alcohol). In particular, there is a long history showing a relation between alcohol and opiate abuse (Siegel, 1986) as well as direct evidence for this relationship (Volpicelli, 1987). Since both alcohol and opiates can involve the abuser with the legal system, a common tendency to abuse them may result from social protest or deviance, rather than addictive tendency. Thus, there is an established path- way of use of addictive substances, progressing to illegal substances, which probably results from the structure of social groups and communication among males from adolescence onward (Kandel & Maloff, 1983).

Send correspondence to: Paul Rozin, Department of Psychology, University OfPennsylvania, 38 I5 Walnut Street, Philadelphia, PA 19104-6196.

This research was supported by a grant from the Whitehall Foundation and the Network on Determinants and Consequences of Health-damaging and Health-promoting Behaviors of the John D. and Catherine T. MacArthur Foundation. We thank Joseph Volpicelli for helpful comments on the manuscript.

81

82 P. ROZIN and C. STOESS

Kagan (1987) has studied people with problem addictions to alcohol, gambling or cigarettes, and finds no uniformity in their scores on six subfactors of the McAndrew Alcoholism scale. He also included a sample of chronic joggers, and found the degree of their jogging negatively related to measures of addiction on this scale. These results are not encouraging for the discovery of an addictive tendency.

The main argument for an addictive tendency comes from theories of addiction, rather than the demonstration of an empirical relation. Insofar as there is a basic under- lying process in addiction that accounts for phenomena such as craving, withdrawal, tolerance and lack of control, it would make sense that individual differences in the operation of this process would lead to variations in addictive tendency. Such a process has been suggested, at the psychological level, in the opponent process theory of addiction (Solomon, 1980). Alternatively, some have proposed that addiction is a product of the characteristics of one or a few neurotransmitter systems. The opiate and dopamine systems have been suggested most often, with the opiate system related to opiates and alcohol (Weisz & Thompson, 1983), and dopamine to the stimulant drugs, such as cocaine and amphetamine (Roberts, Koob, Klonoff, & Fibiger, 1980). It has also been suggested that the opiate system may feed into the dopamine system in the nucleus accumbens, so that at bottom, there might be one dopamine addiction system (Bozarth & Wise, 1986). These speculations have focused on the major, clinically significant addictive substances. Whether the results could be generalized to more minor addictions is yet to be determined.

It seems reasonable, as a first step, and a step that should have preceded the search for the addictive personality, to determine the degree of correlation of different addictions. The problem is difficult to approach, because there is no accepted definition of addiction, and because the term is very widely applied, for example, to include phenomena such as love. Because we are interested in people in general, as opposed to just those for whom addiction has become a serious problem, we have selected a wide range of substances or activities for which there is at least a moderate incidence of addiction. We have defined addiction as the presence of four components: craving, tolerance, withdrawal and lack of control. The paper reports the results of a survey on the extent to which these four components characterize individual relations to ten potentially addicting substances or activities, and the correlations between these items.

METHOD

A survey on body image, addiction, and chocolate preference was distributed to every member of an introductory (270 student registration) psychology class. Students were also given copies of the same questionnaire for their parents; the questionnaire was in an envelope which also included a stamped envelope addressed to the experimenters. Students addressed the outer envelope to their parents, and were encouraged to write a note to their parents at the bottom of the introductory letter accompanying the questionnaire. It is believed that this welcome note increases compliance rates of parents. The students and parents were told that the results from the survey would be shared with the class, and they were. Confidentiality was preserved because each student was arbitrarily assigned a unique code number. Ths number was placed, by the student, on his or her questionnaire form and on those of his or her parents.

There were 573 usable responses, including 130 complete families (mother, father, college student), in which both mother and father were biological parents who had lived continuously with the student. The full sample included 123 sons, 125 daughters, 162

Tendency to become addicted 83

fathers and 163 mothers. Mean ages were 19.3 years for sons, 18.7 for daughters, 49.4 for fathers and 46.0 for mothers. The most substantial racial representation in the sample was 82% white and 14% Asian-American. The major religious affiliations in the sample were 44% Jewish (in accordance with the Jewish representation at the Univer- sity of Pennsylvania), 24% Protestant and 2 1% Catholic.

The part of this survey that is relevant to this study is a set of questions designed to determine whether there is a correlation in degree of addiction to ten different addictive substances or activities. Subjects were given the following instructions (for each of the substances):

“We are interested in the extent to which people can become ‘addicted’ to various activities or substances. For current purposes, we will define addiction as involving craving, withdrawal, loss of control, and tolerance. We define a CRAVING as a strong desire, occurring at least a few times a month. This desire is so strong that it will cause a person to go far out of his or her way to satisfy the craving. For each of the activities and substances below rate your craving on the l-3 scale listed above the list of substances. Write the craving ratings in the first column then return and read about withdrawal.

WITHDRAWAL has to do with the presence of physical discomfort resulting from abstaining from eating, drinking or engaging in the activity of concern for a certain period of time. Using the l-3 scale, please answer the questions about withdrawal in the second column below, and then return to the next paragraph.

LACK OF CONTROL is difficulty in stopping the consumption of the indicated substance or the indicated activity. Using the l-3 scale, please answer the questions about lack of control in the third column below, then return to the next paragraph.

TOLERANCE means that individual doses (portions, levels) of the substance or activity in question produce less of an effect than they used to. Using the l-3 scale, please answer the questions about tolerance in the fourth column below.”

The l-3 scale used was: 3 = strong, definite; 2 = weak, partial; 1 = none. In report- ing scores from this scale, we have subtracted one from each value, so that zero would represent no effect. The total addiction score (sum of the four addiction components for any substance/activity) has a minimum value of zero and a maximum value of eight.

The ten substance/activities listed were (see Table 1): coffee, tea, cola beverages, favorite alcoholic beverage, chocolate, nonchocolate sweets, hot chili pepper on food, cigarettes, gambling and video games. These were arrayed vertically in a matrix with four columns, labelled craving, withdrawal, lack of control and tolerance. Subjects entered a number from 1 to 3 in each cell, according to the instructions listed above.

RESULTS

Addictiveness of d@erent substances and activities The total addictiveness score is highest for chocolate, among all ten items (Table 1).

This is primarily because of the widespread craving for chocolate, such that only 34% of subjects (the lowest number for all items) scored “0,” the minimum addiction score. However, it is also the case that the highest percent of subjects showing substantial addiction scores (> 3 on the O-8 scale) was also for chocolate (2 1.1%; Table 1). This is due primarily to scores on craving and lack of control. The high frequency of use of chocolate by individuals and the high percentage of people who consume chocolate may also contribute to the low level of “0” addiction scores.

Cigarettes generated the strongest addictions among those showing any tendency to


Table I. Degree of addiction to different substances/activities

Substance/activity

Chocolate Coffee Tea Cola Sweets Alcohol Pepper Cigarette Gambling Video games

Addiction overall

Score mean w/o OS % 0 % >3

I .85 2.81 34.0 21.1 I .48 3.07 51.7 17.6 0.64 2.01 68.2 4.2 I.50 2.77 45.9 17.1 1.42 2.41 41.2 14.3 1.14 2.81 55.0 11.6 0.62 2.46 74.9 5.7 0.64 4.06 84.3 8.1 0.38 2.26 83.2 2.9 0.48 2.33 79.2 3.5

II for overall data is 573. For without OS it varies between 90 and 378.

be addicted (scores greater than zero): the mean score for cigarettes among this subset of subjects was 4.06, a full point above any other substance (coffee was next, with 3.07; Table I). Substantial levels of “addiction” (greater than three) were observed in a minimum of 2.9% of subjects for gambling, and a maximum of 21.1% for chocolate (Table 1).

Family resemblance in addiction susceptibility There is a surprisingly small degree of family resemblance in tendency to become

addicted (Table 2). We compare both addiction scores for the ten items, and four scores representing the summed effect, across all ten items, of craving, withdrawal, lack of control and tolerance. With a Bonferroni correction for 14 significance tests on child-midparent (mean of mother and father values) pairings, no item addiction score or addiction component score reaches significance (p < .003); the highest value is only .20). In contrast, for mother-father correlations (14 tests), chili pepper, video games and tolerance show significant effects (p < .OO 1; Table 2). In 12 out of 14 cases, the mother-

Table 3. Family resemblance in addiction scores (Pearson ~j)

Substance/activity Ma-Fa r Midpar-child I

Chocolate .07 .09 Coffee .I2 IO Tea .09 -.I7 Cola .28 .I2 Sweets .06 .oo Alcohol .I6 -.02 Pepper .53*** .20 Cigarette IO .04 Gambling .06 .07 Video games .33*** - .02 Craving .20 .07 Withdrawal .I2 .09 Lack of control .24 .03 Tolerance .45+** - .04

)I = I30 for each category (mother, father, child). ***II < .OO I two-tailed. significant with Bonferroni correction for 14 tests.


Table 3. Correlations between different features of addictions

Craving Withdrawal Lack of control

Withdrawal .59 Lack of control .50 .55 Tolerance .29 .28 .26

n = 573. All correlations significant at p < .OO I.

father correlation is higher than the midparent-child correlation, arguing for effects of assortative mating as more significant than genetic or family influence effects.

Relations of diflerent components of addiction A total score for each of the four components of addiction (craving, lack of control,

withdrawal and tolerance) was created for each subject by adding the scores for that component across all ten substance/activities. All intercorrelations between these four scores are significant at the .OOl level (Table 3). Correlations are particularly robust between withdrawal, craving and lack of control.

Relations of addictions to d.$erent substances/activities Correlations were computed for all pairs of the ten substance/activities, using the

total addiction score for each subject (sum of the craving, withdrawal, lack of control and tolerance scores). The overall impression from this correlation matrix (Table 4) is of positive but modest inter-relations. Of the 45 correlations, 4 1 are positive. Twelve of these are above .20, and 19 are significant at >.OOl (Y 2 .14, given the large n of 573). The largest correlations are eminently understandable. By far the largest is between chocolate and sweets (Y = 53) which, of course, share a sweet component. Next is gambling and video games (Y = .30), which are overlapping activities, followed by chocolate and cola (Y = .28), overlapping in both sweetness and xanthine content, and finally, cigarettes and coffee (Y = .27), two substances often experienced together.

DISCUSSION

The major conclusion of this study is that there is weak but significant evidence (4 1 out of 45 positive correlations) for some common factor in degree of addiction to dif-

Table 4. Correlations in addiction to different substances/activities’

Choc C-Off Tea Cola Sweet Alto Pepp Cigs Gamb

Chocolate Coffee .I2 Tea .I7 .I5 Cola .28 - .06 .08 Sweets .53 .06 .22 .2l Alcohol .I5 .I0 .05 .24 .08 Pepper -.Ol .07 .I4 .I0 .02 .23 Cigarette .03 .27 .07 .08 .04 .22 .22 Gambling .05 .Ol .I0 .I7 .I2 .24 .I7 .03 Video gam -.Ol --.I5 .07 .2l .I I .I9 .08 .oo .30

‘Correlations 1 I4 are significant at the .OO I level, which is the appropriate level with the Bonferroni correction.


ferent substances or activities. The most interesting interpretation of this result is that there is an individual difference in susceptibility to addiction that is, there is a weak general tendency to become addicted. As we pointed out in the introduction, such a claim would have to be established before it would be appropriate to even consider the existence of an addictive personality. However, it is possible that the correlations we report are environmentally, that is, situationally. caused. Some lifestyles, cultural back- grounds and socioeconomic levels may encourage exposure to a whole set of addictive substances or activities. Alternatively, involvement with one addictive substance/activity may increase the probability of involvement with others. Such a relationship has been documented in the use of a variety of legal and illegal pharmacologically active substances. Use of any particular psycho-active drug is correlated with the use of others (Kandel & Maloff, 1983). A common developmental sequence begins with beer or wine, and progresses through cigarettes and hard liquor, then marijuana, and then other illegal drugs (Kandel & Maloff, 1983). Certainly, the coffee/cigarettes combination is a familiar part of American culture, while the common sensory properties of some other items on our list (chocolate, sweets, cola beverages) provide another nonaddictive account for the correlations.

There are also possible sources of weak correlations in the structure of the survey, as it interacts with the interpretations ofthe subject. For example, some subjects may have lower criteria for identifying craving or lack of control than others, which would, of itself, generate some degree of association. Alternatively, the correlations between the components of addiction may have resulted from inferences made by subjects from their frequency of use; if the frequency was high, this may have encouraged high esti- mates for the four addiction components.

We do not have data from our subjects on frequency of use of each of the substances or frequency of engaging in each of the activities. Frequency is presumably related to degree of addiction and might mediate some of the minimal relation between addiction scores for different activities/substances. Ofcourse, even with this data in hand, it would not be clear whether frequency accounted for addiction or vice versa. In any event, any confound with frequency would have the effect of reducing an already very small effect: it is unlikely that incorporating frequency would lead to a better case for a general tendency to become addicted.

It is important to realize that although our survey includes some widely acknowl- edged legal addictive substances, such as alcohol and tobacco, it does not include some of the most addictive, illegal substances, such as opiates and cocaine. It is possible that inclusion of opiates, cocaine and other illegal items in a sample where use of such substances was moderately common might yield results at some variance to those reported here.

In summary, the results of this study support what is at best a very weak tendency to become addicted. In our view, the results should discourage speculation about an addictive personality, at least until more positive data indicating a tendency to become addicted become available.

REFERENCES

Bozarth, M. A.. &Wise, R. A. (1986). Involvement ofthe ventral tegmental dopamine system in opioid and psychomotor stimulant reinforcement. NINA Resc~rch Monogruph.r, 67, l90- 196.

Cox. W. M. (1985). Personality correlates of substance abuse. In M. Galizio & S. A. Maisto (Eds.). Dciermr- nunt.s o/‘,s~rhs~unc~~ uhrrsc. Biologicd, p.s~ddo~icul und cnvironmcniul f2zc.tor.s (pp. 209-246). New York: Plenum.


Kagan, D. M. (1987). Addictive personality factors. JournnlqfPsychology, 121, 533-538. Kandel. D. B., & Maloff, D. R. (1983). Commonalities in drug use: A sociological perspective. In P. K. Lev-

ison, D. R. Gerstein, & D. R. Maloff (Eds.), Commonalities in substance abuse and habitual behavior (pp. 3-28). Lexington: D. C. Heath.

Lang, A. R. (1983). Addictive personality: A viable construct? In P. K. Levison, D. R. Gerstein. & D. R. Maloff (Eds.), Commonalitie.~ in substance abuse and habitual behavior (pp. 157-235). Lexington: D. C. Heath.

Nathan, P. E. ( 1988). The addictive personality is the behavior of the addict. Journal qfConsu/ting and Clin- ical Ps_vchology, 56, I83- 188.

Roberts. D. C., Koob, G. F., Klonoff, P., & Fibiger, H. C. (1980). Extinction and recovery of cocaine self- administration following 6-hydroxydopamine lesions of the nucleus accumbens. Pharmacology, Bio- chemistrv& Behavior, 12, 781-787.

Siegel, S. ( 1986). Alcohol and opiate dependence. Re-evaluation of the Victorian perspective. In H. D. Cap- pell, F. B. Glaser Y. Israel, H. Kalant, W. Schmidt, E. M. Sellers& R. C. Smart (Eds.), Reseurch advances in alcohol and drug problems. volume 9. New York: Plenum.

Solomon, R. L. (1980). The opponent process theory of acquired motivation. American Ps~~chologist, 35, 691-712.

Sutker, P. B., & Allain, A. N. (I 988). Issues in personality conceptualizations of addictive behaviors. Journul ofconsulting and Clinical Psychology, 56, 172- 182.

Volpicelli, J. R. (1987). Uncontrollable events and alcohol drinking. British Journul o/‘Addiction. 82, 385- 396.

Weisz, D. J., & Thompson, R. F. (1983). Endogenous opioids: Brain-behavior relations. In P. K. Levison, D. R. Gerstin. & D. R. Maloff (Eds.). Commonalities in substance abuse and habitual behavior (pp. 297- 322). Lexington: D. C. Heath.

is there a general tendency to become addicted?

Documents