Addictive Behaviors, Vol. 15, pp. 69-72, Printed in the USA. All rights reserved.
1990 Copyright
0306-4603/90 $3.00 + .OO o 1990 Pergamon Press plc
BRIEF REPORT PHYSICAL DEPENDENCE AND ATTRIBUTIONS OF ADDICTION AMONG CIGARETTE SMOKERS DONALD University
S. MARTIN
of New England,
Armidale,
Australia
Abstract - Examination of the attitudes and attributions of cigarette smokers has differentiated smokers who believe their behavior is a “sickness” from those who believe they are “hooked.” Among other things, the hooked smoker, more than the sick one, believes they are addicted and their chances of stopping smoking are poor. If there is a causal association between the attribution of addiction and perceived prospects of change, as this suggests, it could mean treatment and preventative programs stressing the addictive nature of cigarettes may be counterproductive. However, the present study, using a survey of 105 male and female smokers from the general population, suggests the attribution of addiction is related to a smoker’s estimate of their chances of stopping only through a common association that each of these measures has with actual (not necessarily perceived) physical dependence. Caution is needed in the application of cognitive research when related physical measures have not been included in the research design.
In recent work Eiser and associates (e.g., Eiser, 1982; Eiser & van der Pligt, 1986) have examined aspects of the attitudes and attributions of cigarette smokers which could have direct implications for some treatment programs. In particular, they have reported that smokers may perceive their behavior as either a “sickness” or the result of “being hooked.” Among other things, the hooked smoker is more likely than the sick smoker to attribute their behavior to an addiction and to perceive their chances of successfully giving up smoking altogether as being lower. These findings raise the possibility that the “hooked” or “addicted” attribution plays a role in setting a smoker’s perception of their ability to cease, and this in turn may influence the smoker’s responsiveness to change attempts. In fact, Eiser and van der Pligt (1986) have argued that the hooked smoker’s perception of their state “may well be regarded as a major obstacle to behavior change” (p. 14). If correct, this would have important implications for those attempting to modify smokers’ behavior through treatment programs and also for anti-smoking campaigns which stress the addictive nature of cigarettes. Such efforts may be counterproductive if they inadvertently induce a self-attribution of addiction, which carries the conviction for the smoker that they are unable to change. On the other hand, it is possible that the hooked smoker’s perceived difficulty of stopping is not simply a result of the attributional label but reflects a common association between each of these measures and a third, unmeasured variable. It is the contention of the present paper that there is such a third variable and it is actual (not necessarily perceived) physical dependence. The hooked smoker is making their attribution on the basis of an awareness of a dependent state, and it is this latter fact which sets their perceived chances of stopping, not the attributional label. Research which has concentrated only on the cognitive measures has not been able to reveal this possibility. The present study attempted to replicate the association between perceived (attributional)
Requests for reprints should be sent to Dr. D. S. Martin, Department of Psychology, Armidale, N.S.W. 2351, Australia. 69
University of New England,
70
DONALD
S. MARTIN
addiction and perceived probability of cessation but, in addition, it sought information on the physical state of dependence. If the attributional label is to play an important role in setting the person’s perceived chances of stopping smoking, it should be related to that perception independent of the physical state. However, if the association between the attribution and the perceived chances of stopping is spurious, the labeling alone will show no reliable link independent of the physical state. Since the present study was to be conducted outside the laboratory, physical dependence was assessed indirectly using Fagerstrom’s (1978) scale of nicotine tolerance. Despite a recent question mark (McNabb, 1985) this measure has been established and frequently employed as a sound behavioral index of physical dependence (e.g., Fagerstrom, 1983; Fertig, Pomerleau, & Sanders, 1986). Its only limitation in the present context is that it, like the other measures, is a self-report scale, and could be subject to some common error with those measures. However, given the above mentioned validation data, this is likely to be of minimal consequence, and readily outweighed by the advantages of its non-intrusive nature. METHOD
Subjects Fifty-two male and 53 female smokers participated in the study. They were recruited in public places, shopping centers, and in the home. The mean age of the sample was approximately 34 years (SD 12.5). The mean number of cigarettes smoked per day by each person was 20 (SD 8.4). Questionnaire Each person completed a six-page booklet which contained the Fagerstrom Tolerance Questionnaire (TQ), an addiction attribution question from Eiser (1982), a question on perceived chances of stopping smoking altogether, also from Eiser (1982) and three filler questions relating to intention of stopping and the likely response to Government-sponsored campaigns to stop smoking. The addiction question asked “how addicted do you think you are to smoking?’ ’ providing five possible responses from “not at all addicted” to “extremely addicted.” The “chances of stopping altogether” question offered six options from “very poor” to “very good.” The order of questions in the booklet was fixed except for the position of the physiological dependence and attributional addiction questions. Half of the questionnaires presented the TQ scale first and the attribution question last, while the others presented the attribution item first and concluded with the TQ. The two scales were always separated by four questions over three pages. Hence, there was unlikely to be any systematic effect of one item on the other due to position or proximity. RESULTS
Scores were assigned to the three major variables in the following way: A tolerance score on the TQ was assigned following Fagerstrom and Bates (1981) using the Australian Department of Health “Smoker Yield Table” (1984) to score question 2 relative to Australian brands of cigarettes. The attributional addiction question was scored 1 to 5 with the latter figure being assigned to the “extremely addicted” response. Perceived chances of stopping were scored as 1 if the response was “very poor” and so on. Because the pattern of responses revealed by males and females did not differ, and because the position of items on the questionnaire had no effect on responses, all the data will be pooled in the following description.
Dependence
and attributions
71
The mean score on Fagerstrom’s Tolerance Questionnaire was 5.1 (SD 1.8) while the mean addiction attribution was 4.0 (SD 0.9), indicating a “fairly addicted” response. The initial statistical analysis entailed a simple Pearson correlation between the three variables. In replication of the established effect, perceived addiction was moderately associated with perceived chances of stopping smoking (I = - .353, p < .Ol). However, it also correlated with physical dependence as assessed by the TQ (r = .453, p < .Ol), which in turn showed an association with the perceived chances of stopping (r = - .445, p < .Ol). To separate the common variance shared among the three measures, two partial correlations were calculated. This provided information on the possible causal links among the variables (c.f. Blalock, 1962). The partial correlations revealed an association between physical dependence and for the attribution perceived chances of stopping (r = - .34, p < .Ol) controlling measure. When physical dependence was controlled, the attribution measure was only minimally, and not significantly, related to the perceived chances of stopping (r = - .19, N. S .). These figures indicate that the original simple correlation between attributed addiction and perceived chances of stopping primarily reflected a common association between these measures and physical dependence. DISCUSSION
The partial correlations indicate that within the present sample, ascription of the “addicted” label to oneself was not a sufficient reason to believe one’s chances of stopping smoking were reduced. The self-reported level of addiction bore no reliable association with the perceived chances of stopping, independent of the common association with actual physical dependence. Instead, the perceived chances of stopping reliably reflected the physical dependence. Given the complexity of the Fagerstrom measure, and following Blalock (e.g., 1962) we can suggest with some confidence that the smokers’ awareness of their dependent state brought about their poorer perceived chances of stopping. Members of the present group apparently were quite able to appreciate their own state of dependence and, regardless of whether they labeled it addiction or not, they made estimates of their chances of stopping which reflected that state. Moreover, these estimates probably were quite accurate since Fagerstrom and Bates (1981) have already reported a correlation between TQ score and outcome of cessation attempts. It is possible, of course, that some higher level, still unmeasured, variable may mediate all the relationships revealed here and this may be revealed in future research. The relevant variable may be behavioral, physical, or cognitive in nature and hence can only be revealed by continuing the growing trend of research into the interactive effects of relevant causal factors. In the meantime, the results reinforce the calls which are made from time to time for more integration between the psychological and biological approaches to understanding addictive behaviors (e.g., Kozlowski & Herman, 1984). To study one element alone, without consideration of the other, is to run the risk of misinterpreting the role of the highlighted variable in the behavior being studied. The present study has shown that while self-attributions of addiction are associated with a perception of reduced chance of change, they are unlikely to be the cause of that perception. The state of physical dependence at this stage appears to be a more reliable determinant. The present findings reinforce the views of those who have expressed caution (e.g., Fincham, 1983) regarding overzealous application of attribution principles to clinical fields. There is no doubt of the value of knowing more about patients’ attitudes and attributions regarding their condition, but such knowledge must be as thorough as possible before it is
72
DONALD
S. MARTIN
used to modify treatment and prevention programs. In the present case, to direct attention to changing the attribution in the belief that it played an important role in setting the perceived chances of stopping smoking, would be a mistake. It is clear that the attribution itself is, in fact, unlikely to be causal in this way. Finally, the present study employed a cross-sectional design, and its conclusions must be constrained by the limits imposed by that method. It also focused on perceived chances of stopping, which represents an early stage in the quitting process.The next step in the research program is a longitudinal study to follow the process and attendant attributions over a greater time period.
REFERENCES Australian Government Department of Health. (1984). Smoke yield table. Canberra: Author. Blalock, H.M. (1962). Four-variable causal models and partial correlations. American Journal of Sociology, 68, 182-194. Eiser, J.R. (1982). Addiction as attribution: Cognitive processes in giving up smoking. In J. R. Eiser (Ed.), Social psychology and behavioral medicine (pp. 281-299). Wiley: New York. Eiser, J.R., & van der Pligt, J. (1986). “Sick” or “hooked”: Smokers’ nerceutions of their addiction. Addictive I . Behaviors, 11, ll-15. Fagerstrom, K.-O. (1978). Measuring degree of physical dependence to tobacco smoking with reference to individualization of treatment. Addictive Behaviors, 3. 235-241. Fagerstrom, K.-O. (1983). Tolerance, withdrawal and dependence on tobacco and smoking termination. International Review of Applied Psychology, 32, 29-52. Fagerstrom, K.-O., & Bates, S. (1981). Compensation and effective smoking by different nicotine dependent smokers. Addictive Behaviors, 6, 331-336. Fertig, J.B., Pomerleau, O.F., & Sanders, B. (1986). Nicotine-produced antinociception in minimally deprived smokers and ex-smokers. Addictive Behaviors, 11, 239-248. Fincham, F.D. (1983). Clinical applications of attribution theory: Problems and prospects. In M. Hewston (Ed.) Attribution theory: Social andfunctional extensions (pp. 187-203). Oxford: Blackwell. Kozlowski, L.T., Herman, C.P. (1984). The interaction of psychosocial and biological determinants of tobacco use: More on the boundary model. Journal of Applied Social Psychology, 14, 244-256. McNabb, M.E. (1985). Nicotine tolerance questionnaire scores and plasma nicotine: No correlation. Addictive Behaviors, 10, 329-332.
1990 Copyright
0306-4603/90 $3.00 + .OO o 1990 Pergamon Press plc
BRIEF REPORT PHYSICAL DEPENDENCE AND ATTRIBUTIONS OF ADDICTION AMONG CIGARETTE SMOKERS DONALD University
S. MARTIN
of New England,
Armidale,
Australia
Abstract - Examination of the attitudes and attributions of cigarette smokers has differentiated smokers who believe their behavior is a “sickness” from those who believe they are “hooked.” Among other things, the hooked smoker, more than the sick one, believes they are addicted and their chances of stopping smoking are poor. If there is a causal association between the attribution of addiction and perceived prospects of change, as this suggests, it could mean treatment and preventative programs stressing the addictive nature of cigarettes may be counterproductive. However, the present study, using a survey of 105 male and female smokers from the general population, suggests the attribution of addiction is related to a smoker’s estimate of their chances of stopping only through a common association that each of these measures has with actual (not necessarily perceived) physical dependence. Caution is needed in the application of cognitive research when related physical measures have not been included in the research design.
In recent work Eiser and associates (e.g., Eiser, 1982; Eiser & van der Pligt, 1986) have examined aspects of the attitudes and attributions of cigarette smokers which could have direct implications for some treatment programs. In particular, they have reported that smokers may perceive their behavior as either a “sickness” or the result of “being hooked.” Among other things, the hooked smoker is more likely than the sick smoker to attribute their behavior to an addiction and to perceive their chances of successfully giving up smoking altogether as being lower. These findings raise the possibility that the “hooked” or “addicted” attribution plays a role in setting a smoker’s perception of their ability to cease, and this in turn may influence the smoker’s responsiveness to change attempts. In fact, Eiser and van der Pligt (1986) have argued that the hooked smoker’s perception of their state “may well be regarded as a major obstacle to behavior change” (p. 14). If correct, this would have important implications for those attempting to modify smokers’ behavior through treatment programs and also for anti-smoking campaigns which stress the addictive nature of cigarettes. Such efforts may be counterproductive if they inadvertently induce a self-attribution of addiction, which carries the conviction for the smoker that they are unable to change. On the other hand, it is possible that the hooked smoker’s perceived difficulty of stopping is not simply a result of the attributional label but reflects a common association between each of these measures and a third, unmeasured variable. It is the contention of the present paper that there is such a third variable and it is actual (not necessarily perceived) physical dependence. The hooked smoker is making their attribution on the basis of an awareness of a dependent state, and it is this latter fact which sets their perceived chances of stopping, not the attributional label. Research which has concentrated only on the cognitive measures has not been able to reveal this possibility. The present study attempted to replicate the association between perceived (attributional)
Requests for reprints should be sent to Dr. D. S. Martin, Department of Psychology, Armidale, N.S.W. 2351, Australia. 69
University of New England,
70
DONALD
S. MARTIN
addiction and perceived probability of cessation but, in addition, it sought information on the physical state of dependence. If the attributional label is to play an important role in setting the person’s perceived chances of stopping smoking, it should be related to that perception independent of the physical state. However, if the association between the attribution and the perceived chances of stopping is spurious, the labeling alone will show no reliable link independent of the physical state. Since the present study was to be conducted outside the laboratory, physical dependence was assessed indirectly using Fagerstrom’s (1978) scale of nicotine tolerance. Despite a recent question mark (McNabb, 1985) this measure has been established and frequently employed as a sound behavioral index of physical dependence (e.g., Fagerstrom, 1983; Fertig, Pomerleau, & Sanders, 1986). Its only limitation in the present context is that it, like the other measures, is a self-report scale, and could be subject to some common error with those measures. However, given the above mentioned validation data, this is likely to be of minimal consequence, and readily outweighed by the advantages of its non-intrusive nature. METHOD
Subjects Fifty-two male and 53 female smokers participated in the study. They were recruited in public places, shopping centers, and in the home. The mean age of the sample was approximately 34 years (SD 12.5). The mean number of cigarettes smoked per day by each person was 20 (SD 8.4). Questionnaire Each person completed a six-page booklet which contained the Fagerstrom Tolerance Questionnaire (TQ), an addiction attribution question from Eiser (1982), a question on perceived chances of stopping smoking altogether, also from Eiser (1982) and three filler questions relating to intention of stopping and the likely response to Government-sponsored campaigns to stop smoking. The addiction question asked “how addicted do you think you are to smoking?’ ’ providing five possible responses from “not at all addicted” to “extremely addicted.” The “chances of stopping altogether” question offered six options from “very poor” to “very good.” The order of questions in the booklet was fixed except for the position of the physiological dependence and attributional addiction questions. Half of the questionnaires presented the TQ scale first and the attribution question last, while the others presented the attribution item first and concluded with the TQ. The two scales were always separated by four questions over three pages. Hence, there was unlikely to be any systematic effect of one item on the other due to position or proximity. RESULTS
Scores were assigned to the three major variables in the following way: A tolerance score on the TQ was assigned following Fagerstrom and Bates (1981) using the Australian Department of Health “Smoker Yield Table” (1984) to score question 2 relative to Australian brands of cigarettes. The attributional addiction question was scored 1 to 5 with the latter figure being assigned to the “extremely addicted” response. Perceived chances of stopping were scored as 1 if the response was “very poor” and so on. Because the pattern of responses revealed by males and females did not differ, and because the position of items on the questionnaire had no effect on responses, all the data will be pooled in the following description.
Dependence
and attributions
71
The mean score on Fagerstrom’s Tolerance Questionnaire was 5.1 (SD 1.8) while the mean addiction attribution was 4.0 (SD 0.9), indicating a “fairly addicted” response. The initial statistical analysis entailed a simple Pearson correlation between the three variables. In replication of the established effect, perceived addiction was moderately associated with perceived chances of stopping smoking (I = - .353, p < .Ol). However, it also correlated with physical dependence as assessed by the TQ (r = .453, p < .Ol), which in turn showed an association with the perceived chances of stopping (r = - .445, p < .Ol). To separate the common variance shared among the three measures, two partial correlations were calculated. This provided information on the possible causal links among the variables (c.f. Blalock, 1962). The partial correlations revealed an association between physical dependence and for the attribution perceived chances of stopping (r = - .34, p < .Ol) controlling measure. When physical dependence was controlled, the attribution measure was only minimally, and not significantly, related to the perceived chances of stopping (r = - .19, N. S .). These figures indicate that the original simple correlation between attributed addiction and perceived chances of stopping primarily reflected a common association between these measures and physical dependence. DISCUSSION
The partial correlations indicate that within the present sample, ascription of the “addicted” label to oneself was not a sufficient reason to believe one’s chances of stopping smoking were reduced. The self-reported level of addiction bore no reliable association with the perceived chances of stopping, independent of the common association with actual physical dependence. Instead, the perceived chances of stopping reliably reflected the physical dependence. Given the complexity of the Fagerstrom measure, and following Blalock (e.g., 1962) we can suggest with some confidence that the smokers’ awareness of their dependent state brought about their poorer perceived chances of stopping. Members of the present group apparently were quite able to appreciate their own state of dependence and, regardless of whether they labeled it addiction or not, they made estimates of their chances of stopping which reflected that state. Moreover, these estimates probably were quite accurate since Fagerstrom and Bates (1981) have already reported a correlation between TQ score and outcome of cessation attempts. It is possible, of course, that some higher level, still unmeasured, variable may mediate all the relationships revealed here and this may be revealed in future research. The relevant variable may be behavioral, physical, or cognitive in nature and hence can only be revealed by continuing the growing trend of research into the interactive effects of relevant causal factors. In the meantime, the results reinforce the calls which are made from time to time for more integration between the psychological and biological approaches to understanding addictive behaviors (e.g., Kozlowski & Herman, 1984). To study one element alone, without consideration of the other, is to run the risk of misinterpreting the role of the highlighted variable in the behavior being studied. The present study has shown that while self-attributions of addiction are associated with a perception of reduced chance of change, they are unlikely to be the cause of that perception. The state of physical dependence at this stage appears to be a more reliable determinant. The present findings reinforce the views of those who have expressed caution (e.g., Fincham, 1983) regarding overzealous application of attribution principles to clinical fields. There is no doubt of the value of knowing more about patients’ attitudes and attributions regarding their condition, but such knowledge must be as thorough as possible before it is
72
DONALD
S. MARTIN
used to modify treatment and prevention programs. In the present case, to direct attention to changing the attribution in the belief that it played an important role in setting the perceived chances of stopping smoking, would be a mistake. It is clear that the attribution itself is, in fact, unlikely to be causal in this way. Finally, the present study employed a cross-sectional design, and its conclusions must be constrained by the limits imposed by that method. It also focused on perceived chances of stopping, which represents an early stage in the quitting process.The next step in the research program is a longitudinal study to follow the process and attendant attributions over a greater time period.
REFERENCES Australian Government Department of Health. (1984). Smoke yield table. Canberra: Author. Blalock, H.M. (1962). Four-variable causal models and partial correlations. American Journal of Sociology, 68, 182-194. Eiser, J.R. (1982). Addiction as attribution: Cognitive processes in giving up smoking. In J. R. Eiser (Ed.), Social psychology and behavioral medicine (pp. 281-299). Wiley: New York. Eiser, J.R., & van der Pligt, J. (1986). “Sick” or “hooked”: Smokers’ nerceutions of their addiction. Addictive I . Behaviors, 11, ll-15. Fagerstrom, K.-O. (1978). Measuring degree of physical dependence to tobacco smoking with reference to individualization of treatment. Addictive Behaviors, 3. 235-241. Fagerstrom, K.-O. (1983). Tolerance, withdrawal and dependence on tobacco and smoking termination. International Review of Applied Psychology, 32, 29-52. Fagerstrom, K.-O., & Bates, S. (1981). Compensation and effective smoking by different nicotine dependent smokers. Addictive Behaviors, 6, 331-336. Fertig, J.B., Pomerleau, O.F., & Sanders, B. (1986). Nicotine-produced antinociception in minimally deprived smokers and ex-smokers. Addictive Behaviors, 11, 239-248. Fincham, F.D. (1983). Clinical applications of attribution theory: Problems and prospects. In M. Hewston (Ed.) Attribution theory: Social andfunctional extensions (pp. 187-203). Oxford: Blackwell. Kozlowski, L.T., Herman, C.P. (1984). The interaction of psychosocial and biological determinants of tobacco use: More on the boundary model. Journal of Applied Social Psychology, 14, 244-256. McNabb, M.E. (1985). Nicotine tolerance questionnaire scores and plasma nicotine: No correlation. Addictive Behaviors, 10, 329-332.
