Tobacco Documents Online

I I I I I I I I I I I I I I I I I Running addiction: measurement and associated psychological characteristics CAROL LEE CHAPMAN*, JOHN M. DE CASTRO From the Department of Psychology, Georgia State University, Atlanta Ga., U.S.A. Rt.nners frequently become addicted to run- ning and continue to run even when it is detrimental to their health. In the present study a paper and pencil test to measure run- ning addiction, the running addiction scale (RAS) was constructed and evaluated and used to investigate the psychological correlates of running addiction. The RAS, symptom check list, locus of control scale, commitment to run- ning scale (CR), and a questionnaire about the individual's running habits and degree of ad- diction were administered to 32 male and 15 female runners. The results suggest that the RAS is reliable and valid, correlating with self- rated addiction. The CR score was correlated with self-rated addiction for males but not for females suggesting that the RAS and CR meas- ure different characteristics and that for fe- male runners commitment to running can oc- cur without addiction. Large gender differ- ences were found in the correlations between the RAS, CR, run frequency, and run duration with the personality, mood, and locus of con- trol scores. Running addiction was found to be associated with high frequency of running and with positive personality characteristics but not with mood enhancement, while the du- ration of running was found to be associated *The results of the present study were in part der- ived from a senior honors thesis performed by the first author who is commited but not addicted to running under the direction of the second author who is both commited and addicted. The significant contributions of Dr. James L. Pate and Dr. Walter A. Peiper are gratefully acknowledged. Appreciation is also expressed to Drs. William P. Morgan and Michael L. Sachs for reviewing and commenting upon the initial version of the running addiction scale and to Elizabeth P. Shuler for helpful com- ments on the manuscript. This research was sup- ported in part by Grant RO1-DK39881-01A2 from the National Institute of Diabetes and Digestive and Kidney Diseases. with mood enhancement. These results sug- gest that the benefits of running to mood may be obtained without addiction. [J Sports Med Phys Fitness 1990;30:283-90]. Key words: Running - Psychology Many individuals have taken up running for fitness and enjoyment. Some, however, appear to have become addicted, continu- ing to run even to the detriment of their social lives, work, or even health.' Even after having been advised by a physician not to run, some cbntinue, resulting in serious injuries to their ankles, knees, hamstrings, etc. Such self injurious be- havior only occurs in a small number of runners and only after they've been run- ning for months or years. As Sachs and Pargman z point out, running addiction, like other addictions, should be consi- dered a process rather than a condition, developing slowly. For example, a runner writes; "I have run more than 25,000 miles... I won't be able to run very many more. The problem is a degenerating hig,... I have always been plagued by injuries... But should I regret the running I have done, since it probably led to the injury I have now? I don't think so. I was able to run for 20 years, and the only harm to come from it is that I can't run now. It's as if the harm caused by a drug addiction was that the addict eventually wasn't able to use drugs anymore. ... would I do the Vol. 30, No. 3 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS 283 t

I I I I I I I I I I I I I I I I I same high mileage I did before? ...-how could I have resisted the temptation?" (Spillman, p. 10).3 Currently there is no suitable scale to measure running addiction. Carmack and Martens ^ developed a scale, "Feelings about running", to measure commitment to running (CR). However, Thaxton 5 found that runners' CR scores were not signifi- cantly correlated with their self-reported addiction scores. This suggested that com- mitment and addiction are different processes and that the CR is not a suita- ble instrument for the measurement of running addiction. Conceptually, addic- tion to running may be viewed as a process which compells an individual to run in spite of obstacles and results in physical and psychological symptoms when withdrawn. Commitment, on the other hand, is an intention to continue running and results in feelings of satisfac- tion, enjoyment, and accomplishment. Ad- diction may be viewed as a push or a drive while commitment can be viewed as a pull or an incentive. The present investigation attempted to investigate this distinction and conceptualization and to develop a reliable and valid paper and pencil meas- ure of running addiction as distinct from commitmemt. The second objective of the present study was to identify the personality characteristics that are associated with running addiction. These then may serve as predictors and might possibly be used to identify individuals who are likely to be- come addicted to running. A number of studies have investigated the personality characteristics of runners 6-9 but have not separated addicted from non-addicted runners. Sachs z attempted to assess per- sonality traits in relation to running ad- diction. He correlated the scores of 100 male runners on introversion-extro- version, neuroticism-stability, mood states, trait anxiety, and locus of control with their commitment to running (CR) scale scores. He found "... only very modest correlations between any of the traits and level of addiction" (p. 121) and concluded that the measured traits were not good predictors of addiction. These findings, however, could have occurred because the CR measures commitment rather than addiction. It is possible that with a stronger measure of addiction per- sonality traits may be identified that are clearly associated with running addiction. In the present study, the psychological characteristics of runners were assessed with a symptom check list (SCL-90-R) and Levenson's locus of control scale. In ad- dition, the runners were asked to rate their own level of addiction and their lev- el of discomfort when a run is missed, to fill out the commitment to running (CR) scale, and to report the characteristics of their running, usual frequency, time and distance. All of these scores and ratings were intercorrelated and correlated with the running addiction scale. Materials and methods Subjects and procedure Subjects were obtained by telephone contact with 56 members of the Georgia State University road runners club. They were asked to participate in an investiga- tion of runner's attitudes and personali- ties by completeing anonymously four questionnaires that would be mailed to them. Of the 52 members who agreed to participate, 47, 32 males (24 to 57 years of age) and 17 females (25 to 59 years of age), completed and returned the question- naires. Participants answered a brief ques- tionnaire asking their age, seac, length of average run in minutes, number of runs per week, number of years running, whether or not they experienced discom- fort when a scheduled run was missed, and a self-rating on a 1 to 7 scale of how addicted to running they consider them- selves to be. They also completed a run- ning addiction scale, a commitment to running scale, the symptom check listi`? 284 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS September 1990E'p Q.7 W ~ ~ ~ _ . ~

I I I I I I I I I I I TABLE I.-Running addiction scale (RAS) items, corrected item total correlations (CITC) and correla- tions with self-rated addiction. CITC Self-rated addiction Items (included in RAS) I run on a regular basis .62 .35 If the weather is too cold, too hot, or too windy, I will not run that day .69 .59 I would not reschedule activities with my friends in order to run .36 .54 I have stopped running for a period of at least a week for a reason other than an injury .41 .20 I would run with intense pain .51 .32 I have not spent much money on running literature or equip- ment and accessories .46 .29 If there were another way to maintain my present physical fitness, I would not run anymore .50 .37 After I run I"feel" better .48 .19 I would continue to run while an injury healed .50 .34 On some days, even though I do not feel like running, I do anyway .41 .27 I feel that I need to run at least once every day .54 .41 Weak items (not included in the final RAS) My family and friends have rarely complained about my run- ning .13 .07 If I miss a scheduled run, I feel guilty .10 .07 I run to stay in shape .21 .08 I think of running when I am doing other activities, such as worling .26 .21 When I miss a scheduled run, I feel that my level of fitness drops .18 .16 I would not miss an appointment in order to run .06 .07 If during a run, I have to walk, for even a short period of ti- me, I feel that I have let myself down -.12 -.04 (SCL-90-R), and a locus of control scale as- sembled in a packet in random order. Materials the revised RAS. The RAS items were then randomly mixed with the items from the Carmack and Martens' ° commitment to running (CR) scale to produce the final form for administration. Running addiction (RAS) and Commit- ment to running (CR) scales. - The RAS was constructed by writting 18 statements regarding the characteristics associated with running addiction (i.e., withdrawal ef- fects, feelings of a need to run, etc.). Par- ticipants were asked to respond using a five point likert scale from strongly agree to strongly disagree. The statements were administered to five runners in a pilot study were evaluated by two prominent researchers in the area. Their comments and suggestions were incorporated into Personality and locus of control scales. - The symptom check list, SCL-90-R 10 was used. It is a 90 item self-rep9rt sym- ptom inventory measuring nine symptom dimensions; somatization, obsessive- compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxi- ety, paranoid ideation, and psychoticism. Levenson's tt locus of control scale was al- so employed. It measures three control dimensions: powerful others, chance, and internal. Vol. 30, No. 3 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS 285

I I I I I I I I I I I I I I I I TABLE II.-Means and standard deviations o f the RAS, CR, SCL-90-R, locus o f control, and running characteristics of male and female runners. Mean Ma les Fema le s Standard Standard M ean deviation deviation Running addiction scale (RAS) 34.8 7.9 33.4 7.3 Commitment to running (CR) 45.8 8.2 42.4 8.4 Self-rated addiction 3.6 1.6 3.8 1.7 # of runs per week 3.9 1.5* 3.9 1.3 Duration of average run 33.1 9.1 31.0 8.7 Symptom check list (SCL-90-R) Somatization 49.2 8.9 50.2 8.8 Obsessive-compulsive 56.2 9.0* 52.8 9.9 Interpersonal sensitivity 53.6 9.4* 59.1 8.9 Depression 53.9 10.5 57.3 10.9 Anxiety 52.8 9.0 54.0 10.2 Hostility 50.9 7.7* 57.2 10.2 Phobic anxiety 49.4 5.9 49.8 7.7 Paranoid ideation 51.5 8.3 53.5 9.7 Psychoticism 51.0 9.2 54.9 7.0 Locus of control Powerful others 14.8 4.3 16.6 3.9 Chance 13.8 3.0* 16.2 4.0 Internal 28.6 2.6* 25.6 4.5 *Indicates a significant, t-test, p<.05. difference between the male and female runners. TABLE III.-Correlations between the running addiction scale (RAS) scores, commitment to running (CR) scores, sel f-rated addiction and discomfort. 4, RAS CR score .799* Self-rated addiction .664* Discomfort .475* Males 'Indicates p<.05. Results Running addiction scale construction and evaluation Each of the 18 items in the RAS was as- sessed for reliability by calculating the corrected item total correlation. Each was, further, assessed for validity by cal- culating each item's correlation with the self-rated addiction to running. These are presented in table I. Eleven of the items were judged to have satisfactory correct- ed item total correlations (range: .36 to .69) Females CR RAS CR .644 * .762* .753* .246 .504* .391 -.126 and validity correlations (range: .19 to .59) while 7 were judged as unsatisfactory. The final 11 item RAS had a coeffl,'cient alpha of .82, an acceptable level of reliability. The mean scores and their standard deviations of the male and female runners on the final 11-item RAS and all of the other measures are presented in table II. There were no significant differences amoung the males and females on the RAS, CR or running characteristics. There were, however, considerable individual differences between the runners with the 1 286 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS September 1990

I I I I I I I I I I I I I I 1 I TABLE IV.-Correlations between the running addiction scale (RAS) scores, the commitment to run- ning (CR) scores, and the personality characteristics of male and female runners. RAS CR Runs/week Run duration M F M F M F M F RAS score .80* .64* .69* .61* .41* .34 CR score .80* .64* .59* .14 .16 .28 Symptom check list (SCL-90-R) Somatization .30 .40 .37* -.04 .23 .27 .02 .08 Obsessive-compulsive -.30 .00 -.26* -.12 -.33 -.30 -.41* .16 Interpersonal sensitivity -.43* .16 -.23 -.20 -.41* 12 -.52* -.02 Depression -.35 .12 -.18 -.11 -.24 .12 -.57* .15 Anxiety -.31 .05 -.11 -.26 -.30 .11 -.64* -.03 Hostility -.08 .02 -.04 -.43 -.04 .21 -.38* -.17 Phobic anxiety -.41* -.12 -.44* -.19 -.52* .34 -.32 -.01 Paranoid ideation -.10 .13 -.07 -.16 -.19 .16 -.39* -.25 Psychoticism -.40* .26 -.27 -.10 - .36* .16 -.34 -.18 Locus of control Powerful others .09 -.08 .07 -.14 .05 -.07 -.41* -.69* Chance -.13 -.01 -.17 .02 .03 -.16 -.25 -.63* Internal .24 .37 .19 .17 -.04 .00 .38* .14 *Indtcates p<.05. number of runs per week ranging from 1 to 7 and the duration of the average run ranging from 15 to 60 minutes. The RAS scores (range: 22 to 53), CR scores (range: 19 to 59), and self-rated addiction scores (range: 1 to 7) also indicated large in- dividual differences. The scores on the new 11 item RAS were correlated with the CR scale scores, the self-rated addiction scores and the scores for discomfort experienced when a sched- uled run was missed. These correlations, presented in table III, suggest that the RAS is a valid measure of running addic- tion since it correlates strongly with self- rated addiction and moderately with dis- comfort for both male and female runners. They also suggest that the RAS and the CR measure different characteristics. The CR did not significantly correlate with self- rated addiction in females while the RAS did (z=2.00, p<.05). Table III also suggests that there is a sex difference in the relationship between ad- diction and commitment. For the males, the RAS and CR correlate similarly with the self-rated addiction and discomfort scores. However, the correlations for the females show a different pattern; the RAS is significantly correlated with self-rated addiction while the CR is not. A similar pattern occurs with the correlations be- tween the discomfort scores and the RAS and CR. This suggests that addiction and commitment may be equivalent for males but constitute different dimensions of run- ning for females. Running characteristics, personality and locus o f control correlates The mean scores and their standard deviations of the male and female runners on the nine scales from the SCL-90-R and the three locus of control scales as well as the running characteristics are present- ed in table II. The symptom check list scales use T scores with a normative mean of 50 and standard deviation of 10. The males are above the norm for obsessive- compulsive tendencies and significantly higher than the females; t(45)=2.26, p <.05. The females appear to be above the norm in hostility and interpersonal sen- Vol. 30, No. 3 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS 287 ~

I I I I I I I I I I I I I I I sitivity and are significantly higher than the males; t(45)=2.37, p<.05, and t(45) = 2.03, p<.05, respectively. The locus of control scale results indicate that both the male and female runners are high in internal control but the males are signifi- cantly higher; t(45)=2.30, p<.05. Both groups have low scores for powerful others and chance locus of control with the male significantly lower than the fe- males on chance locus of control, t(45)=2.92, p<.05. In order to investigate the psychologi- cal characteristics associated with run- ning and addition, correlations were cal- culated between the scores from the sym- ptom check list, locus of control scale, RAS and CR and the RAS and CR scores, the frequency of running and the duration of the average run. These correlations are presented in table IV. It is clear from this table that the frequency of running and the duration of the average run are as- sociated with the RAS score. Addicted run- ners tend to run frequently and for long durations. The CR score correlated signifi- cantly with run frequency for the male but not the female runners and not with run duration for either sex, further suggest- ing that commitment and addiction to run- ning are different especially for females. The large male/female difference is again apparent in the correlations be- tween the running characteristics, the RAS, and the CR scores and the personal- ity characteristics ascertained from the symptom check list. There were no signifi- cant correlation for females while there many for the males. Also apparent is the fact that the pattern of correlations be- tween the RAS score and the personality characteristics is parallelled by the run frequency correlations with these same characteristics. A striking characteristic of table IV is the fact that the correlations indicate a relationship betweem positive personali- ty characteristics and addiction, high fre- quency and long duration running. In par- ticular, high addiction scores and high run frequencies are associated with low inter- personal sensitivity, phobic anxiety, and psychoticism scores. Run duration was al- so negatively associated with inrerper- sonale sensitivity. Interestingly, unlike run frequency and addiction, run duration was also negatively correlated with depression, anxiety, hostility, paranoid ideation, and obsessive-compulsive ten- dencies. Additionally, run duration was as- sociated positively with an internal locus of control and negatively with a powerful others locus of control. Thus there was a clear positive relationship between run- ning addiction, run frequency, and run du- ration and psychological health. Discussion The results suggest that the revised, 11- item, RAS is a reliable and valid measure of addiction to running. Internal reliabil- ity is indicated by an acceptable coeffi- cient alpha. The validity of the scale is in- dicated by its face validity and its strong correlations with self-rated addiction and discomfortewhen a scheduled run is missed. The RAS also shows discriminant validity by correlating differently than the CR scores with self-rated addiction and run frequency for female runners. Those females who considered themselves ad- dicted to running scored high on the RAS but not on the CR. For these runners com- mitment and addiction appear to be differ- ent processes. It would thus appear that the RAS may be a better way to measure true addiction than to measure enjoyment and commitment. The RAS then appears to be a reliable and valid tool in the inves- tigation of running addiction. A striking characteristic of the present findings is the marked difference between the results for male and female runners. The RAS and CR for males had similar patterns of correlations with the other variable while for females distinctly different patterns emerged. This suggests 288 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS September 1990 I

I I I I I I I ~ 1 I I I I I I I that for males commitment and addiction are closely related phenomena, while fe- males may be committed to running without becomming addicted. Further, there were no significant correlations for female runners between the personality or mood scores and the RAS, CR, or running characteristics, while for males there were many. This suggests that for females, run- ning is not associated with personality or mood. The only exception was in the case of the locus of control scores and their correlations with the duration of the aver- age run for female runners. Although chance and powerful others locus of con- trol scores were low for runners in gener- al, female runners who on the average run long distance tend to see their lives as less controlled by either chance or powerful others. Male runners who run for long du- rations also tend to see themselves as less controlled by powerful others, but unlike the females tend to be more internally con- trolled. Why there are such marked differences between male and female runners was not addressed in the present study. The males and females in the study were not signifi- cantly different in addiction or commit- ment to running, frequency or duration of running, or on many of the psychological characteristics. Thus, there were no obvi- ous differences between the characteris- tics of the male and female runners that might account for the observed gender differences. The present findings, then, suggest that the males and females may be running for different reasons and ob- taining different benefits. For male runners the RAS and CR scores had similar patterns of correla- tions. The RAS, however, consistently had stronger correlations with the other vari- ables than the CR. Consistent with the present findings, Sachs (1981) found only very modest correlations between CR scores and personality scores in male run- ners. He suggested, however, that the per- sonality traits he chose to measure were not good predictors of addiction. The present finding suggest that the trouble was that the CR score was not a good measure of addiction and thus produced only weak correlations. The present results suggest that the ad- dicted male runner is one who runs fre- quently and for long distances, who is low in interpersonal sensitivity and psychoti- cism and who has low levels of phobic anxiety. This implies that running addic- tion is associated with positive, not nega- tive, psychological characteristics. It is no wonder, then, that otherwise completely well adjusted individuals can become strongly addicted to running. The running characteristics also demon- strated an interesting pattern of correla- tions with the psychological characteris- tics of male runners. The frequency of run- ning correlated highly with addiction and had a pattern of correlations with the psy- chological characteristics that was very similar to that of the RAS scores. Run du- ration, on the other hand, had a somewhat different pattern of correlations. Although there was some similarity in the correla- tions, run duration appears to be more highly related to elevation of mood than is addiction or run frequency. In particu- lar, there were strong negative correla- tions between run duration and depres- sion, anxiety, and hostility. These results are compatible with previous findings. Morgan and Costill (1972) found that marathon runners were below normal in anxiety. Morgan and Pollock (1977) found that world class distance runners were low in tension, depression, fatigue, and confusion. This suggests that the positive mood enhancement benefits of running may be obtained without addiction. Mood enhancement may occur with long dis- tance running independent of running fre- quency. This suggests the prescription of few but long runs to improve mood but to avoid addiction. In sum, the present findings indicate that the RAS is a reliable and valid scale Vol. 30, No. 3 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS 289 I

I I I I I I I I I I I I I I I for the measurement of running addiction. They suggest that male and female runners differ markedly in their reasons for run- ning and the benefits obtained from run- ning, They suggest that running addiction is associated with high frequency of run- ning and with positive psychological characteristics but not with mood enhance- ment. Finally they indicate that the dura- tion of running is associated with mood en- hancement, implying that the benefits to mood may be obtained without addiction. References 1. Morgan W. Negative addiction in runners. Phys Sports Med 1979;7:57-70. 2. Sachs M. Rtutning addiction. In: Sachs MH, Sachs ML, eds. Psychology of running. Cham- paign IL: Human Kinetics Publishers, 1981. 3. Spillman D. Goodbye to running. The Wingfoot. Newsletter of the Atlanta Track Club 1987; 61(8):10. 4. Carmack M, Martens R. Measuring commitment to running, a survey of runners' attitudes and mental states. J Sports Psych 1979;1:25-42. 5. Thaxton L. Physiological and psychological ef- fects of short-term exercise addiction on habitual runners. J Sports Psych 1982;4:73-80. 6. Clitsome T, Kostrubala T. A psychological study of 100 marathoners using the Myers-Briggs type indicator and demographic data. Ann NY Acad Sci 1977:301:1010-9. 7. Gontang A, Clitsome T, Kostrubala T. A psycho- logical study of 50 sub 3 hours marathoners. Ann NY Acad Sci 1977;301:1020-8. 8. Morgan W, Pollock M, Psychological characteris- tics of the elite distance runner. Ann NY Acad Sci 1977;301:382-403. 9. Morgan W, Costill D. Psychological characteris- tics of marathon runners. J Sports Med Phys Fit 1972;12:42-6. 10. Derogatis L. The SCL-90-R. Baltimore, MD: Clin- ical Psychometric Research, 1976. 11. Levenson H. Multidimensional locus of control in psychiatric patients. J Con Clin Psych 1973;41:397r404. Address reprint requests to: J. M. de Castro - Depart- ment of Psychology, Georgia State University, Univer- sity Plaza - Atlanta, GA 30303 (USA). . 1 290 THE JOURNAL OF SPORTS MEDICINE AND PHYSICAL FITNESS