What Is the Myth?^
The major thrust of our article was based on Schachter's (1982) argument that inferences drawn from studies of single attempts to quit smoking are not directly relevant to lifetime quitting. The data Schachter reported, however, did not directly support this important point. He found that rates of retrospective self-reports of self-quitting over the lifetime were much higher than rates reported in evaluations of one-shot clinic quitting. Were these rates different because self-quitters were in some way different from clinic quitters, because self-quitting is different from clinic quitting, or because multiple attempts to quit (during a lifetime) are more likely to result in a successful quit than is a single attempt?
We hoped to provide answers to these questions by making a fair comparison between self-quitting and clinic quitting. We did this by providing data on single-quit attempts for self-quitters and comparing their success rates with archival reports of single-quit attempts of clinic quitters. There were no differences in success rates of self-quitters and clinic quitters. We thus inferred that the differences Schachter found were attributable to his asking about lifetime quitting and not to self-selection or to the relative effectiveness of self- and clinic quitting. This is what we were debunkingthe myth that self-quitting is more successful than clinic quitting. When a fair comparison was made, it was not!
We were also surprised that Schacter and Forgays and Forgays ignore evidence that we reported regarding the dynamics of the quitting process and their implication for understanding lifetime quitting. This included process data not usually reported in smoking cessation studies: evidence for late quits (after expected quit dates), late relapses (after six months of abstinence), and the influence of previous quit attempts on the probability of a successful quit.
The comments also ignore our discussions of how quitting is defined and how different definitions can result in very different conclusions about how successful people are at quitting smoking. Finally, they ignore the fact that we confirmed Schachter's (1977) argument that smoking rate should predict success in quitting even though his own 1982 data did not.
Who Is the Population?^
The only substantive criticism provided by the comments has to do with whether our (Cohen et al., 1989) samples were representative of our target population. Forgays and Forgays (1990) argue that the samples in Rzewnicki and Forgays's (1987) article and Schachter's (1982) article provided excellent representations of their target populations. For example, they say that the strength of Schachter's findings is that he obtained smoking cessation data from a very large proportion (98%) of the membership of two separate populations [italics added] (p. 1387). We doubt that Schachter's intent was to generalize his result only to members of the Columbia University psychology department and entrepreneurs in Amaganset, or that Rzewnicki and Forgays's (1987) intent was to generalize their results only to members of the University of Vermont psychology department. These were narrow, selected samples in a demographic sense relative to either the population of self-quitters or to the 589 persons sampled in our studies.
Our (Cohen et al., 1989) intent was to sample the population of smokers who were about to make a serious attempt to quit smoking by themselves. Although there is no national survey data on the characteristics of persons who are about to make serious attempts to quit smoking, there is reason to believe that our samples are representative of this target population. Specifically, the characteristics of our samples were consistent with characteristics of persons concerned with health and health practices. For example, our samples had fewer men than women. This is consistent with data reporting that men have shown a greater failure to follow standard health practices, report symptoms, and visit physicians than have women (Lewis & Lewis, 1977; Nathanson, 1977; Verbrugge, 1985; Waldron, 1976). Similarly, our samples tended to have levels of education and income higher than the national average, and persons with higher socioeconomic status are also more likely to show concern with health and to follow health practices (Bunker, Gomby, & Kehrer, 1989). In short, those concerned with health and health practices resemble our subjects. We expected that these were the same people who would be making serious attempts to quit smoking.
Our studies, of course were neither perfect nor definitive. Our own concerns with their limitations (including sample representativeness) were addressed in some detail in the discussion of the original article (Cohen et al., 1989). We believe the questions raised in our article will be best answered by a combination of data from national surveys, retrospective reports of lifetime quitting, and long-term prospective follow-ups of persons attempting to quit smoking. Each method has its own costs and benefits but we hope that when viewed together they will provide an understanding of the dynamic process of quitting smoking.
Procedural and Statistical Details^
We were surprised by the commenters' misinterpretation of the method section of our article (Cohen et al., 1989). First, the article was not a review of published research in this area, but rather a report of original data. Data from 7 of the 10 studies had not appeared in any form elsewhere, and only select data were used from the 3 studies with previously published quitting rates (see the section entitled Study Descriptions, p. 1356 for selection criteria). The references to the studies included in the Cohen et al., (1989) article were references to detailed descriptions of the procedures (p. 1356), and were not intended as references to original sources of data. Second, we are hard-pressed to understand what methodological and statistical clarifications Schachter (1990) could be concerned with. The designs, sample characteristics, and procedures of the studies were covered reasonably comprehensively on pages 13561358, and the statistics in our article were primarily means and medians with an occasional chi-square. Most disturbing, however, is Schachter's implication that we intentionally gerrymandered the sample in the Marlatt, Curry, and Gordon (1988) study so that the results would fit our hypothesis. We clearly indicated the criteria for subject inclusion in our analyses (p. 1356). One of these criteria was that subjects enter studies (baseline data) prior to their attempts to quit smoking. This is what made these studies prospective! The Method section of the Marlatt et al., article indicated that only 69 of their subjects entered the study prior to quitting; these were the 69 who were included in our analyses. It is obvious that mixing a sample of persons who have already quit smoking with one of persons about to attempt to quit would inflate long-term quit rates relative to samples (like those in our analysis) that include only those recruited prior to their quit attempts.
Forgays and Forgays (1990) appear to think that the essence of our article was the reestablishment of the stereotype of the hopelessness of self-cure of cigarette smoking (p. 1388). This is just absurd. What were our messages? (a) It is hard to quit smoking and stay off cigarettes, whether in a clinic or by oneself; (b) quitting is a dynamic process not a dichotomous event; and (c) failures to quit in the past do not influence success rates on future attempts (Cohen et al., 1989, pp. 13621364). Both optimism and pessimism are generated by our results. However, our role as scientists is to accurately report the data we have collected. Neither our understanding of the addiction process nor our ability to aid persons in quitting are facilitated by maintaining a myth that it is easier to quit smoking than it actually is.
Bunker, J. P., Gomby, D. S., & Kehrer, B. H. (Eds.). (1989). Pathways to health: The role of social factors. Menlo Park, CA: Henry J. Kaiser Family Foundation. [Context Link]
Cohen, S., Lichtenstein, E., Prochaska, J. D., Rossi, J. S., Gritz, E. R., Carr, C. R., Orleans, C. T., Schoenbach, V. J., Biener, L., Abrams, D., DiClemente, C., Curry, S., Marlatt, G. A., Cummings, K. M., Emont, S. L., Giovino, G., & Ossip-Klein, D. (1989). Debunking myths about self-quitting: Evidence from 10 prospective studies of persons who attempted to quit smoking by themselves. American Psychologist, 44, 13551365. [Context Link]
Forgays, D. G., & Forgays, D. K. (1990). Seventeen authors in search of a myth: A reply to Cohen et al. American Psychologist, 45, 13871389. Ovid Full Text Library Holdings [Context Link]
Lewis, C. E., & Lewis, M. A. (1977). The potential impact of sexual equality on health. The New England Journal of Medicine, 32, 523. [Context Link]
Marlatt, G. A., Curry, S., & Gordon, J. R. (1988). A longitudinal analysis of unaided smoking cessation. Journal of Consulting and Clinical Psychology, 56, 715720. [Context Link]
Nathanson, C. A. (1977). Sex, illness and medical care. Social Science and Medicine, 11, 1325. [Context Link]
Rzewnicki, R., & Forgays, D. G. (1987). Recidivism and self-cure of smoking and obesity: An attempt to replicate. American Psychologist, 42, 97100. [Context Link]
Schachter, S. (1977). Nicotine regulation in heavy and light smokers. Journal of Experimental Psychology: General, 106, 512. [Context Link]
Schachter, S. (1982). Recidivism and self-cure of smoking and obesity. American Psychologist, 37, 436444. [Context Link]
Schachter, S. (1990). A reply to Cohen et al. American Psychologist, 45, 13891390. Ovid Full Text Library Holdings [Context Link]
Verbrugge, L. M. (1985). Gender and health: An update on hypotheses and evidence. Journal of Health and Social Behavior, 26, 156182. [Context Link]
Waldron, I. (1976). Why do women live longer than men? Social Science and Medicine, 10, 349362. [Context Link]
Accession Number: 00000487-199012000-00028