Back to Journal Cover Page
Volume 5, Number 16
Submitted: July 17, 2000
Resubmitted: July 26, 2000
Accepted: July 27, 2000
Publication date: August 3, 2000
Saint Francis Xavier University
Saint Francis Xavier University
Two studies were conducted to test the hypothesis that fame induces chronic self-awareness, which in turn is avoided through strategic self-destruction (Schaller, 1997). In Study 1 analyses of Ernest Hemingway's short stories indicated increased use of first-person voice after he attained celebrity. In Study 2 private and public self-consciousness, self-esteem and self-reported alcohol use were assessed in relatively well-known (famous) and less well-known students and faculty/staff members. Relatively well-known participants scored significantly higher on all measures; also, a significant fame by self-esteem interaction on alcohol consumption was found, suggesting that self-esteem might play a moderating role between self-awareness and self-abuse. Overall, these results offer additional empirical support for Schaller's hypothesis.
For most people fame represents a very attractive and desirable status--so many privileges can be associated with celebrity. And yet, anecdotal sources (e.g., magazines, TV programs and the internet) regularly report stories of addiction and suicide in famous people, suggesting that this status might also be linked to painful and negative consequences. More serious sources (Ludwig, 1995) indicate that among eminent people, 26% experience alcohol-related problems during their life-times compared to 14% in the general population; also, the rate of suicide for famous people is 4.4% compared to 1.0% for non-famous individuals. A popular explanation proposes that fame is somehow stressful, but this hardly constitutes a satisfying answer since it is vague and incomplete: why would fame be unpleasant, and why would many celebrities engage in self-destruction?
An intriguing possibility has been recently raised by Schaller (1997). Famous people are constantly being observed and scrutinized by others; as a result, they very frequently become the object of their own attention. In this perspective, fame can be seen as a situational factor inducing chronic self-awareness. Celebrities are obviously set apart from the majority of the less-eminent population. The group that contains persons who have obtained fame status is much smaller than the population to which the rest of society belongs. Investigations have shown that people tend to develop an intensifying focus on the self as the actual size of their group decreases (e.g., Diener, Lush, DeFour & Flax, 1980). Thus, fame would represent a special, extreme case of individuation, causing heightened self-awareness.
In a state of self-awareness a person will actively look at and examine any self-aspect that is most salient at the moment (Buss, 1980). Self-awareness theory (Duval & Wicklund, 1972; Wicklund, 1975, 1978) suggests that high levels of self-focus are usually painful. More precisely, self-awareness following failure feedback is associated with negative affect, whereas self-awareness following success feedback is associated with positive affect (Carver & Scheier, 1981; Hull, 1981). The reasoning is that self-directed attention initiates a comparison between the "real" self and the "ideal" self; this process is likely to eventually uncover some perceived personal shortcomings, causing discomfort. Then psychological mechanisms will be activated to strategically eliminate this negative affect. The initial reaction is to simply escape the state of self-awareness by avoiding whatever self-focusing stimuli caused it. If this is impossible or ineffective, the individual will try to reduce the discrepancy between the real self and the ideal self, either by directly modifying the real self or by changing the ideal self. (For a review of evidence, see Carver & Scheier, 1981; Davis & Franzoi, 1991.)
In essence these mechanisms are healthy; however, fame-induced self-awareness is likely to be more intense, and more dangerous coping strategies may be initiated. Temporary self-awareness may focus one's attention on specific aspects of the self, but famous people are presumably in a permanent state of self-awareness. It seems plausible that chronic self-awareness will lead to a more global evaluation of one's self-worth (Gibbons, 1990). And unlike most people, celebrities cannot avoid the source of self-awareness--their omnipresent fans. Moreover, for this very exclusive group of individuals, the strategy of discrepancy reduction is likely to be unsuccessful (Schaller, 1997). The culture of fame often exaggerates expectations beyond a level that can be realistically reached, thus attainment of ideals may be impossible; also, modification of ideals will be difficult because the expectations to which public figures respond are not fully under their control but instead reflect the ideals defined by others.
Some famous individuals will resort to an extreme strategy in order to reduce any negative affect caused by chronic self-observation--drug, and in particular, alcohol, abuse. Alcohol is known to interfere with the cognitive processes underlying self-awareness (Hull, 1981). Put simply, heavy drinking lowers self-focus. This could explain why some celebrities, being unable to change their real and ideal selves or to escape self-awareness, will try to avoid themselves through excessive drinking. Other extreme strategies used to escape chronic self-awareness may include suicide, overeating, hypersexuality, and religion (Baumeister, 1990, 1991).
Schaller (1997) partially tested the hypothesis that fame induces chronic self-awareness, which in turn is avoided through strategic self-destruction, by conducting single-case quantitative studies. He produced three biographical analyses of famous persons known for their self-destructive behaviors: songwriters Kurt Cobain and Cole Porter, and writer John Cheever. Schaller measured self-awareness by calculating the number of first-person singular pronouns found in the songs, short stories or personal letters of these three celebrities. An increase in the use of first-person singular pronouns indicates high self-directed attention, because the more one thinks about oneself, the more one is likely to refer to oneself in one's writings (Davis & Brock, 1975; Carver & Scheier, 1978). Using biographies, Schaller also determined the exact moment these individuals attained fame, and measured Cheever's self-reported alcohol consumption by analyzing his personal letters. As predicted, the onset of fame induced high self-focus. In other words, Cobain, Porter and Cheever began to use significantly more first-person singular pronouns in songs, stories and personal letters following their brush with fame. Also, this onset was significantly related (in Cheever's case) to higher self-reported alcohol use.
In this paper we present two studies designed to expand these preliminary results. In Study 1 we measured self-awareness and self-reported alcohol intake using Schaller's method for one additional celebrity: Nobel Prize winner Ernest Hemingway--a well-known heavy drinker. We postulated that an analysis of this celebrity's writings would show a significant increase of self-awareness and drinking patterns following fame.
In Study 2 we addressed the question of moderating variables. All celebrities do not live miserably in a constant state of drunkenness. How can we explain the fact that some famous people do not find chronic self-awareness to be painful--and consequently do not try to self-escape? Schaller (1997) and Baumeister (1991) suggest that self-esteem might play an important role in that respect (also see Brockner & Wallnau, 1981): well-adjusted celebrities would have a healthy, positive self-esteem. Consequently, and as mentioned previously, the real and ideal selves would nearly coincide and chronic self-observation would not lead to unpleasant effects because very few personal shortcomings would be identified in this state. Therefore, these individuals should not feel the urge to avoid themselves through excessive drinking or drug abuse. We tested this assumption by assessing self-awareness, self-esteem, and self-reported alcohol use with questionnaires in 245 well-known (i.e., relatively famous) and not well-known students and faculty/staff members. In accordance with the Schaller's hypothesis, we predicted that self-focus and alcohol use would be significantly higher in the group of well-known participants. In addition, it was postulated that self-esteem would have a moderating effect on alcohol intake.
The goal of Study 1 was to replicate Schaller's original investigation by conducting one additional single-case quantitative analyses.
STUDY 1 METHOD
Ernest Hemingway was born in 1899. Hemingway's life spread over a sixty-two year span and was quite amazing. Biographies indicate that he became famous in 1929 after the publication of his book Farewell to Arms (Burgess, 1978; Crisman, 1998a): this book was highly praised by critics, and the general public instantly took notice and enjoyed Hemingway's new literary style; that generated much attention on him. Then Hemingway had ups and downs--his fame fluctuated throughout his life. In 1959, he was awarded a Pulitzer Prize for his novel The Old Man and the Sea. The following year he was awarded the Nobel Prize. Hemingway had a well-established and documented reputation of a heavy drinker. He also suffered from severe bouts of depression and committed suicide at his home in 1961.
Procedure and Method
Schaller's original method (1997) was used to determine and compare levels of self-awareness and self-reported alcohol use before and after fame was attained. Hemingway wrote 61 short stories. A chronological list of the Finca Vig collection (1998) was made, and all short stories were sorted into two categories: those written (i.e., published) before 1929 (when Hemingway attained fame) and those written after that date. Hemingway wrote 28 short stories before 1929 and 33 short stories after 1929. To determine levels of self-awareness, each story was identified as being narrated in either the first-person voice (indicating high self-focus) or third-person voice (indicating lower self-focus). The number of stories narrated in the first-person voice was expected to be significantly higher after 1929 than before that date1.
Hemingway's writings also provided the opportunity to examine the relationship between fame and self-abuse. Given the uneven trajectory of Hemingway's fame, a correlational strategy was also used to test the hypothesis that fluctuations of his celebrity over time would be associated with a corresponding variation of self-reported alcohol use. Fame was operationalized as the number of citations to Hemingway in the New York Times Index and the Readers Guide to Periodical Literature. Citations were summed up from these two sources, and a composite quantitative indicator of fame was generated for every year from 1920 to 19612. Alcohol consumption was operationalized as follows: Hemingway's personal letters (published by Barker, 1981) were examined, and whether or not alcohol use was mentioned in each letter was recorded. A typical example would be Hemingway complaining of a massive hangover or stating that he spent the night out drinking. It was assumed that thoughts about drinking, or mentions of having been drinking, would be representative of actual drinking behavior.
Since the addressees might exert an influence on the content and style of Hemingway's letters, analyses focused only on letters he wrote to close friends and longtime correspondents. These were John Dos Passos (13 letters from 1925 to 1949), Charles Scribner (27 letters from 1939 to 1952), and F. Scott Fitzgerald (25 letters from 1925 to 1935). Mentions to alcohol were summed up from these three sources, and a composite quantitative indicator of self-reported alcohol use was generated for every year from 1925 to 1952. Alcohol was mentioned in 23 letters. One first prediction was that the number of letters in which alcohol was referred to would to be significantly higher after 1929 than before that date. Also, since Hemingway's fame varied in time, a correlational strategy was used to verify if fluctuations of his celebrity over time would be associated with a corresponding variation of self-reported alcohol use. Fame was operationalized as described above with the indexes; the indicator of fame was then correlated with self-reported alcohol use in the letters. A positive and significant correlation between fame and self-reported alcohol use was expected3.
STUDY 1 RESULTS
A before/after strategy was used to compare the narrator voice of Hemingway's short stories (indicator of self-awareness) with fame. A significant effect of fame on self-awareness was found, F(l.59) = 10.65, p < 0.05. The first-person voice in Hemingway's short stories significantly increased from 0.233 (before 1929) to 0.452 (after 1929) (see Table 1).
Table 1. Self-Awareness and References to Alcohol before and after Fame.
|Before Fame||After Fame||p|
Note. Hemingway's self-awareness was based on narrator voice in short stories he wrote before and after publication of his book Farewell to Arms. His alcohol use was based on mentions to alcohol in a subset of his personal letters written between 1925 and 1952. See text for details.
T-tests were calculated to determine if Hemingway made significantly more mentions to alcohol in personal letters written once fame was established. No significant effect of fame on self-reported alcohol use was found: Hemingway did not refer to alcohol more frequently in his letters after 1929 (.319, compared to .353 before) (see Table 1). A correlation was calculated to determine the possible existence of a relation between fame and self-reported alcohol use in Hemingway's personal letters. No significant correlation was found between the number of citations in the indexes and mentions to alcohol use in letters, r = 0.13, ns.
STUDY 1 DISCUSSION AND ADDITIONAL ANALYSIS
Overall these results provide additional, although partial, support to Schaller's original findings (1997): levels of self-awareness significantly increased as Hemingway became more famous. Schaller (1997) identified (and eliminated) two plausible alternative causal explanations for the increase of self-awareness following fame in Cobain, Porter and Cheever. Perhaps fame itself did not make Hemingway more self-aware--rather, creative output (i.e., the number of published short stories following fame) may have caused heightened self-focus. Additional analyses indicated that this was unlikely. Aside from his increased fame and self-awareness in years in which his short stories were published, Hemingway's fame (mentions in indexes) nor self-awareness (narrator voice in short stories) was significantly related to the number of short stories he published each year after 1929--r = -.33 (ns) and -.43 (ns) respectively. Also, since chronological age was confounded with fame, maybe Hemingway became increasingly self-aware with age, regardless of fame? This second alternative explanation was tested by numbering his short stories according to the chronological order they were written, and by computing a correlation between these numbers and the measure of self-awareness (first/third person voice in each short story after 1929). A negative (non-significant) correlation of .11 was obtained, clearly indicating that age was not responsible for Hemingway's increased self-focus once fame was attained.
As mentioned previously (also see notes 3), it proved impossible to directly test the hypothesis of a correlation between self-awareness and self-reported alcohol use for Hemingway. However, it was possible to test the hypothesis according to which fame would lead to self-destructive tendencies. This hypothesis was not supported: no relation was observed between fame and mentions to alcohol use in Hemingway's personal letters.
One limitation of Schaller's approach is that it is idiographic--it focuses on changes over time within the same person. This obviously raises concerns about the generality of his, and of our, results. Another limitation is that only famous songwriters or novelists can be studied, since written material is needed in order to come up with an objective evaluation of subjects' levels of self-awareness. One goal of the second study was to go round these limitations by assessing self-awareness and self-reported alcohol use with validated scales in a large group of well-known (i.e., relatively famous) and rather unknown (i.e., obscure) individuals. We assumed that if genuine celebrity creates heightened self-focus and fosters self-abuse, then similar effects, although possibly to a lesser degree, should be observed in a relatively famous population. Also, using a totally different population, as well as different measures, to test Schaller's hypothesis should make it even more plausible if the expected results are found.
A second goal of Study 2 was to examine the role of self-esteem as a moderating variable between chronic self-awareness and self-abuse. Having a positive self-esteem means that one is satisfied with oneself. Consequently, as a result of self-awareness, any comparison made between the real self and the ideal self is likely to lead to positive feelings about oneself. This might explain why some celebrities may not find chronic self-focus to be painful and may not need to escape from the self through alcohol or drug abuse. Recall that in Study 1 a lack of available data made it impossible to directly test the hypothesis of a relation between self-awareness and alcohol self-reported use (only the relation between fame and alcohol self-reported use was examined). Study 2 allowed a direct test of this hypothesis.
STUDY 2 METHOD
One hundred and fifty relatively well-known participants and 95 less known undergraduate students and faculty/staff members at St. Francis Xavier University (Nova Scotia, Canada) participated in Study 2. Subjects were contacted personally, by mail, or by telephone. The Student Telephone Directory was used to obtain individual office locations, mail box addresses, and telephone numbers.
Relatively Famous Students
Well-known participants were socially exposed individuals in the university community; because of their unique position they were easily recognizable by other persons and had achieved relative fame. Five criteria were used for the selection of relatively famous students. The first selection criterion was being a member of the Student Union executive or Student Union sub-executive. Holding a position on these committees puts students in the public eye. More specifically, these students' names, positions held, and pictures are published in the student handbook. The handbook is distributed to every university student. Also, the name, student union position, office number, and telephone number are typed in boldface in the Student Telephone Directory, which is also distributed to every student and faculty/staff member. Thirty-four students were selected using this criterion and participated in the study.
The second criterion for relative fame was being a Residence House President. Each residence house on campus has a president who is responsible for the house activities and finances. These individuals also hold the position of Student Council member. Therefore, house presidents are highly visible and known to all of the students who reside in their house, as well as to the other house presidents and to all those who sit on Student Council. There are twelve residence houses on campus; therefore, twelve students were included from this criterion and participated in the study. Third, any member of the Residence Staff at the university was considered to be well-known on campus. More precisely, Residence Staff members (residence directors and residence assistants) actively interact with all of the individuals who reside in their respective houses. Fifty students were included from this criterion and participated in the study.
A fourth criterion was being a member of the student senate or student board of governors. The names and pictures of the student senators and board of governor members are published in both the university student handbook and the Xaverian Weekly (the student newspaper) at the beginning of the school year. These individuals are well-known to the student population because they are student contact persons for academic and financial concerns. Eight individuals were included from this criterion and participated in the study. Finally, any individual who served as an Off-Campus counsellor was also considered well-known, since he or she, much like House Presidents, is responsible for the availability of activities and awareness of the off-campus students who live in their assigned jurisdiction within the campus area. Each off-campus councilor must visit the houses of the students in their jurisdiction at least once during the year, as well as send out weekly e-mail messages to all of those students. As a result, these persons are highly visible within the student community. Thirteen students were included from this criterion and participated in the study.
Relatively Unknown Students
The selection of the less well-known students involved three basic criteria. First, these participants had to be in their first year of university, to ensure that none had been in attendance at the university for more than five months. The less time one has spent in one location, the less known one is likely to be. Second, mature students (age 30 and more) were excluded as participants. The rational behind this criterion was to ensure that the ages of relatively unknown students were similar to those of typical university students. Third, potential participants had to be registered as full-time students rather than part-time students. This criterion was used in order make sure that each less well-known individual was spending the same amount of time in classes interacting with other students. The relatively unknown students were selected from introductory psychology classes at St. Francis Xavier University. A total of 67 students were included in this group and participated in the study.
Well-Known Faculty/Staff Members
Three criteria were used for the selection of the well-known faculty/staff members. First, any individual who held the position of Chair of his or her academic department was included as a highly visible faculty member. Second, any person who held a position in which he or she was the head of a specific department, program or office within the university, was also included as an highly visible and well-known staff member. For example, Manager of Residence and Food Services, Communications Officer, Student Union Administrator, and the three university physicians, were all considered candidates from this criterion. Third, the university president, three vice presidents (Academic, Administrative and Advancement), three deans (Arts, Science, and of Students), as well as the assistant to the deans, were included. Thirty-three participants were invited to be part of this group; all accepted to participate in the study.
Relatively Unknown Faculty/Staff Members
The less well-known participants were randomly selected among the other faculty or staff members. Twenty-eight participants were included in this last group.
The variables being assessed in Study 2 were self-awareness, self-esteem, and alcohol use. The Self-Consciousness Scale (SCS) was used to tap levels of private and public self-focus. The SCS has been traditionally used to measure the disposition to spend more or less time in the state of self-awareness (Fenigstein, Scheier & Buss, 1975). Self-consciousness is considered to be a stable personality trait, but we assumed that it would be sensitive to situational and chronic self-awareness. The scale consists of 17 items; 10 items assess private self-consciousness--the disposition to engage in fantasies about oneself, to look inside the self and to inspect motives, moods, and mental processes. A sample item is "I'm generally attentive to my inner feelings". Seven items measure public self-consciousness, the disposition to be aware of, and concerned with, the impression one makes on others, one's behavior in social settings, and one's physical appearance. A sample item is "I'm usually aware of my appearance". Participants are asked to evaluate each item on Likert-type scales ranging from extremely uncharacteristic of myself (0) to extremely characteristic of myself (4). A total score is obtained by adding up each rating.
Rosenberg's Self-Esteem Scale (SES; Rosenberg, 1965) was chosen to evaluate participants' self-esteem. The SES represents a popular and validated measure, consisting of 10 short questions related to one's self-worth. Participants are asked to evaluate each item on Likert-type scales ranging from Strongly agree (1) to Strongly disagree (5). A total score is obtained by adding up each rating. A sample item is "I feel I have a number of good qualities". A simple and direct scale was used to assess alcohol use (Weinberger & Barthalornew, 1996). The measure consists of three questions about drinking patterns. Two questions are about the frequency of drinking, one for beer and wine, one for hard liquor. Participants are given the option of choosing a number between 1 (Never used) and 9 (More than once per day); number 5 states "More than once per month". Scores obtained for each question are combined to form one composite quantitative indicator of frequency of drinking. The third question assesses the number of drinks a participant typically has when he or she does consume alcohol.
A single package, which included the three measures as well as basic information about sex and age, was distributed in person to the relatively well-known students and faculty/staff members, and to the less well-known faculty/staff members. Participants were asked to fill-in the questionnaires immediately. Less well-known students filled-in the questionnaires in class and received credits for their participation in the study. The packages were labeled in such a way to distinguish if they had been completed by (1) relatively well-known students, (2) relatively well-known faculty/staff members, (3) less well-known students, or (4) less well-known faculty/staff members. The independent variables consisted of status of the student or faculty/staff member (well-known vs not well-known) and self-esteem. The dependent variables were private/public self-consciousness and reported alcohol intake.
Study 2 tested three main hypotheses. First, self-consciousness was expected to be higher in the well-known group of students and faculty/staff members than in the relatively unknown group of students and faculty/staff members. There were no specific predictions concerning private and public self-focus. A second hypothesis stated that well-known individuals would consume significantly more alcohol in comparison to relatively unknown individuals. Finally, it was hypothesized that self-esteem would have a moderating effect on alcohol intake. Alcohol intake was predicted to be significantly lower in well-known students and faculty/staff members with high self-esteem; also, an overall negative correlation between self-esteem and alcohol intake was expected.
STUDY 2 RESULTS AND DISCUSSION
Preliminary analysis revealed no important gender differences. Data from males and females were thus combined in the analyses reported below. Also, the pattern of results was similar across the two samples of well-known/not well-known students and faculty/staff members. Therefore, results were combined. A multivariate analysis of variance was used to test the first and second hypotheses. A significant main effect was observed for relative fame on all dependent variables: private self-consciousness, F (1.182) = 10.12, p < .01; public self-consciousness, F (1.182) = 5.37, p < .03; frequency of alcohol use, F (1.182) = 10.15, p < .01; and quantity of alcohol intake, F (1.182) = 10.94, p < .01. Well-known individuals scored significantly higher on all measures in comparison with relatively unknown individuals (see Table 2). Since private and public self-consciousness were highly correlated, we combined the two scores into one index--total self-consciousness. A significant positive correlation between total self-consciousness and quantity of alcohol consumed (0.43) was revealed, as well as between frequency of alcohol use and quantity of alcohol consumed (0.49) (all p's < .05). Also, a significant negative correlation was observed between total self-consciousness and self-esteem (-0.54).
Table 2. Private and Public Self-Consciousness, Self-Esteem, and Self-Reported Alcohol Use in Well-Known and Not Well-Known Participants.
|Frequency of Alcohol Use||9.97||8.50|
|Quantity of Alcohol Consumed||6.73||4.95|
Note. All p's < .05
A regression analysis was used to assess the significance of the interaction between relative fame and self-esteem on alcohol consumption. The categorical variable "fame" was effect coded to make it appropriate for use in a regression analysis. Well-known participants were assigned a value of 1 whereas not well-known individuals were given a value of -1; the self-esteem scores were converted to z-scores before computing the interaction term. Results showed a significant fame by self-esteem interaction on both frequency of alcohol use, F(1.188) = 9.71, p < .01, and quantity of alcohol consumed, F(1. 188) = 10.17, p < .01. Simple effects analyses (presented in Table 3) indicated that in well-known participants, the negative correlations between both frequency and quantity of alcohol consumption and self-esteem were significant, whereas only the correlation between quantity of alcohol consumed and self-esteem was significant in not well-known participants. Actually, the relation between self-esteem and quantity of alcohol consumed explained only 4.8% of the variance in the not well-known group, whereas it explained 21% of the variance in the well-known group.
Table 3. Correlations between Self-Esteem and Frequency and Quantity of Alcohol Consumption in Well-Known and not Well-Known Individuals
|Frequency of Alcohol Use||-0.37*||-0.06|
|Quantity of Alcohol Consumed||-0.46*||-0.22*|
* p < .05
The following exploratory analyses were also performed. A significant main effect of sex was found in regards to frequency of alcohol use, F (1.182) = 6.58, p < .02, and quantity of alcohol use, F (1.182) = 9.22, p < .01. Males were found to consume alcohol more frequently and in greater quantity than females. Although there were no specific previsions regarding students and faculty/staff members, a number of significant differences in the means were revealed (all p's < .05). The mean public self-consciousness score for students (x = 19.17) was higher than that of faculty/staff members (x = 17.13). The mean score for students on private self- consciousness (x = 24.40) was higher than that of faculty/staff members (x = 21.87). Furthermore, the mean score for quantity of alcohol consumption by students (x = 6.53) was higher than the mean score for faculty/staff members (x = 4.36).
Results of Study 2 were consistent with the primary hypothesis, which stated that self-consciousness would be higher in the well-known group of participants, in comparison to the relatively unknown group of participants. The well-known students and faculty/staff members scored significantly higher on both private and public self-consciousness. This finding provides further empirical support to the hypothesis proposed by Schaller (1997), according to which celebrity (or relative fame) will lead to heightened, and perhaps chronic, levels of self-awareness.
Results of Study 2 also supported the second hypothesis stating that the well-known individuals would consume significantly more alcohol than the less well-known individuals. The frequency of alcohol use and the quantity of alcohol consumed were both significantly higher in the well-known group of students and faculty/staff members. It can be proposed, as the hypothesis would suggest, that relatively well-known participants drank more because of heightened self-awareness. In other words, relative fame would lead to increase self-focus, and then high self-focus would motivate alcohol intake. We suggest that the existence of a positive correlation between self-awareness and alcohol consumption supports this view. Results showed that in our sample the more participants were self-aware the more they drank. (Note that it would not make sense to say that the more participants were drinking the more they were self-aware, since alcohol use has been shown to decrease self-awareness [Hull, 1981]). This is consistent with Schaller's position (1997), according to which alcohol can be strategically used to lower self-focus.
In theory it is assumed that high self-focus generated negative emotions, and that, in turn, these painful emotions were dealt with through alcohol use. No evidence can directly support this assumption, but the fact that a negative correlation exists between self-consciousness and self-esteem at least partially supports the idea that high self-awareness can generate negative affects. Results show that highly self-aware participants had a tendency to evaluate themselves negatively (clearly an uncomfortable state), and vice-versa. Although the possibility that low self-esteem could foster high self-focus cannot be ruled out, the correlation seems to support the notion that chronic self-awareness can lead to global self-evaluation, which then would shed light on perceived shortcomings (Gibbons, 1990).
Study 2 also disclosed results in favor of the third hypothesis, according to which self-esteem would represent a moderating factor in the equation. In our sample alcohol self-reported use was inversely related to self-esteem--participants with lower self-esteem reported higher alcohol consumption, and vice-versa. This would support the assumption that self-esteem moderates the effects of SA on self-abuse: fame would bring high self-focus; this would lead to painful emotions only if self-esteem is low, and then to higher alcohol use as a means of decreasing self-awareness. Self-focus would not represent a painful experience in participants with high self-esteem, and alcohol use would be lower.
Taken together, we believe that the results of both studies presented in this paper lend additional support to Schaller's hypothesis (1997) of a relation between fame, chronic self-awareness, and self-abuse. Celebrity did exacerbate Hemingway's self-awareness; but some effects were not significant, and the rather crude operationalizations of self-focus and alcohol use might partially be responsible. The fact that we were able to eliminate other alternative causal explanations for Hemingway (creative achievement and age) gives even more weight to the results. In short, replicating some of Schaller's results with one additional case yields yet greater support to the hypothesis.
Study 2 explored the hypothesis by using a totally new paradigm--different measures and participants--, and all expected relationships were observed. This gives much more generality to the results. It is important to realize that rather strong effects were obtained in a population of relatively famous individuals. We believe that even stronger effects would be found in a population of really famous people. It must be noted here that some uncontrolled differences between our well-known and unknown groups might affect levels of self-consciousness and alcohol intake. For instance, the well-known students were more advanced in their studies than the unknown students (all of whom were in their first year). This might have created a "renown/maturity" confound. Future studies should give more attention to this problem and also exclude some third variable alternatives, such as talent, socioeconomic level, past use of alcohol, and stress.
The demonstration that self-esteem represents a moderating variable between self-awareness and self-destruction in our sample constitutes an exciting extension to Schaller's work. Obviously, additional replication across more cases is needed, as well as more investigations in large groups of genuinely famous individuals4.
Sex differences in drinking patterns were found in Study 2. Overall, male participants reported using alcohol more frequently and in larger quantity than females. This observation raises an intriguing possibility. Alcohol abuse might represent the favored self-destructive behavior in famous males, whereas famous females might choose to engage in different types of self-escape--reclusion, anorexia nervosa, bulimia, overeating, promiscuity, religion, or suicide (Baumeister, 1991). Investigation into the possible sex differences in self-destructive strategies used by celebrities would certainly be interesting.
Many celebrities engage in self-destructive behavior. Just very recently, actor David Strickland committed suicide by drug overdose in an hotel room. Actors Charlie Sheen and Robert Downey Jr. repeatedly had problems with substance abuse and the law. These are just examples--the list could be endless. Understanding the psychological mechanisms underlying these self-destructive behaviors certainly represents a legitimate theoretical avenue. More practical applications could also be considered: an increased knowledge of the psychology of fame will allow us to identify individuals at risk and develop adequate preventive programs to support these celebrities.
1. It proved impossible to directly test the hypothesis of a link between self-awareness and fame. To calculate a meaningful correlation, we needed self-awareness data (evaluated with the short stories) from 1925 to 1952. However, Hemingway did not publish short stories between 1941 and 1956.
2. These data are consistent with the suggestion that Hemingway's fame fluctuated throughout his life. There were only two citations before 1929; after one first spike of fame between 1929 and 1939 (20 citations), a second one occurred in 1940-41 (36 citations). Between 1941 and 1950 only 15 citations were made. Then, in 1950-51 a second, more important spike was observed (55 citations).
3. (See note 1.) It also proved impossible to directly test the hypothesis of a link between self-awareness and self-reported alcohol use. Again, to calculate a meaningful correlation, we needed self-awareness data from 1925 to 1952. However, Hemingway did not publish short stories between 1941 and 1956.
4. Investigating the psychology of fame in genuine celebrities might be problematic. In a preliminary phase of Study 2, the participation of more than 100 famous Hollywood actors was solicited. The questionnaires were sent to these actors' agents, together with a letter in which we explained the nature of the study and insisted on the confidentiality of the results. Only one actor responded favorably to the study.
Barker, C. (1981). Ernest Hemingway, selected letters, 1917-1961. New York : Scribner.
Baumeister, R. F. (1990). "Suicide as escape from self." Psychological Review, 97:90-113.
Baumeister, R. F. (1991). Escaping the self. New York: Basic Books.
Brockner, J., and L. B. Wallnau (1981). "Self-esteem, anxiety, and the avoidance of self-focused attention." Journal of Research in Personality, 15:277-291.
Burgess, A. (1978). Emest Hemingway and his world. NewYork: Charles Scribner's.
Buss, A. H. (1980). Self-consciousness and social anxiety. SanFrancisco: Freeman.
Carver, C. S. and M. F. Scheier (1978). "Self-focusing effects of dispositional self-consciousness, mirror presence, and audience presence." Journal of Personality and Social Psychology, 36:324-332.
Carver, C. S. and M. F. Scheier (1981). Attention and self-regulation: A control theory approach to human behavior. New York: Springer-Verlag.
Crisman, D. (Producer). (1998a). Biography: Ernest Hemingway - Wrestling with life (AAE-14366). Idaho Publications. Academic Press.
Davis, D. and T. C. Brock (1975). "Use of first person pronouns as a function of increased objective self-awareness and prior feedback." Journal of Experimental Social Psychology, 11:381-388.
Davis, M.H. and S. L. Franzoi (1991). "Self-awareness and self-consciousness." In V.J. Derlega, B.A. Winstead, & W.H. Jones (Eds.), Personality: Contempory theory and research. pp. 314.347. Chicago: Nelson-Hall.
Diener, E., R. Lush, D. DeFour and R. Flax (1980). "Deindividuation: Effects of group size, density, number of observers, and group member similarity on self-consciousness and disinhibited behavior." Journal of Personality and Social Psychology, 39(3):449-459.
Duval, S. and R. A. Wicklund (1972). A theory of objective self-awareness. New York:
Fenigstein, A., M. Scheier and A. Buss (1975). "Public and private self-consciousness: Assessment and theory." Journal of Consulting and Clinical Psychology, 36:1,241-1,250.
Finca Vig Edition (1998). The Complete Short Stories of Ernest Hemingway. New York: Scribner.
Gibbons, F. X. (1990). "Self-attention and behavior: A review and theoretical update." In M.P. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 23). New York: Academic Press. Pp. 249-303.
Hull, G. (1981). "A self-awareness model of the causes and effects of alcohol assumption." Journal of Abnormal PVchology, 90:586-600.
Ludwig, A. M. (1995). The price of greatness. New York: The Guilford Press.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.
Schaller, M. (1997). "The psychological consequences of fame: Three tests of the self-consciousness hypothesis." Journal of Personality, 65:291-309.
Weinberger, D. and K. Bartholomew (1996). "Social-emotional adjustment and patterns of alcohol use among young adults." Journal of Personality, 64(2):495-527.
We thank Georgina Petrie, Daniel Voyer and Chris Davis for their help during data collection and analysis. Requests for reprints should be sent to Alain Morin, 823 Nouvelle-Orléans, Ste-Foy (Québec), CANADA G1X-3J4.
Alain Morin received his Ph.D in psychology from Laval University (Quebec, Canada) in 1992. He has been teaching and doing research on self-awareness at many different French and English speaking institutions. He is currently at Acadia University (NS, Canada). Web page: www.upc.qc.ca/~morina. Mailing address: 823 Nouvelle-Orléans, Ste-Foy, (Québec) G1X-3J4 CANADA. E-mail: email@example.com.
Lisa Craig was an undergraduate student in Psychology at St. Francis Xavier University (NS, Canada) when the present research was being conducted.
Back to Journal Cover Page