Adolescent attributes contributing to the imposter phenomenon☆
Abstract
Personality traits, self-perceptions, beliefs, and feelings that accompany the Imposter Phenomenon (IP) have been identified in adults but little research with adolescents has been reported. The present research describes data from a sample of 11th and 12th graders in order to examine predictors of IP among adolescents. The participants completed measures of the IP, global self-worth, social support and self-concept. Gender differences were found in correlations between IP and Parent Support, Classmate Support, and Teacher Support. Multiple regression analyses found significant predictors of IP scores for females to be Friend Support, Classmate Support and Dependability. Only Friend Support significantly predicted IP scores for males. The unique variance explained (UVE) by each of the variables was fairly modest, suggesting that the variables are explaining a considerable amount of the same portion of the variance, particularly for females.
Introduction
The Imposter Phenomenon (IP) was first identified in the clinical work of Clance and Imes (1978) as an internal experience of fraudulence despite one's external success. Harvey and Katz (1985) defined IP as “a psychological syndrome or pattern based on intense, secret feelings of fraudulence in the face of achievement tasks and situations.” Initially identified in high achieving women, it was later found to exist among men as well, although the research is contradictory as to whether or not it is more prevalent among women (Harvey, 1981; Topping & Kimmel, 1985; Cozzarelli & Major, 1990; King & Cooley, 1995).
Those experiencing IP attribute success and achievement to external sources such as luck, charm, good social contacts, and hard work rather than to internal qualities of intelligence and ability. Imposters do not believe that their successes are warranted; they feel inadequate and undeserving. Despite academic achievements, advanced degrees, and professional status, imposters believe that they have deceived others regarding their intellect and must not allow others to discover their lack of competence (Clance & Imes, 1978; Harvey & Katz, 1985).
When faced with a project (and therefore an obligation to perform), imposters experience the “Imposter Cycle” (Clance & Imes, 1978). This cycle begins with worry, self-doubt, and intense fears of discovery which then leads to either procrastination or over-preparation. This is typically followed by the project's success, but the success is then attributed to some external cause. The procrastination or over-preparation, for the imposter, reinforces her belief that success was not authentic.
Research has supported the theoretical concept of IP. Although its features include introversion (Holmes, 1986), self-monitoring and anxiety (Topping & Kimmel, 1985), the IP is a distinct construct. It includes the fear of evaluation, the fear of not being able to repeat success and the fear of not being able to perform as well as others (Edwards, Zeichner, Lawler, & Kowalski, 1984). For imposters, the development of self-perceptions becomes self-critical, discounting and inaccurate. Imposters develop self-perceptions of incompetence and fraudulence despite objective information to the contrary (Clance & Imes, 1978; Harvey & Katz, 1985). Other characteristics shown in the literature include poor interpersonal flexibility (Hays & Davis, 1993), fear of success as well as fear of failure (Fried-Buchalter, 1992), defensiveness (Beard, 1990), and low self-esteem (Harvey, 1981; Topping & Kimmel, 1985; Kertay, 1991).
The IP was first conceptualized as a gender issue. Clance and Imes (1978) asserted that IP was primarily limited to women due to cultural sex-role expectations. However, subsequent research showed that IP feelings were shared similarly by both males and females (Harvey, 1981; Lawler, 1984; Cromwell, 1990).
What does appear to be different among male and female imposters is personal style. Beard (1990) found that while both genders doubted their abilities to succeed, their interpersonal styles were quite different. In her sample, female imposters were described as self-protective, cautious, lacking in social initiative, and avoiding of interpersonal connections, whereas male imposters were more accepting of social interactions and did not tend to withdraw from social contact.
Clance and Imes (1978) identified low self-acceptance in their original conceptualization of the phenomenon. Later studies supported this observation by identifying the negative relationship between self-esteem and IP (Harvey, 1981; Topping & Kimmel, 1985; Kertay, 1991). In addition to these studies, Eschbach (1990) discovered that both self-satisfaction and self-concept had strong negative correlations with IP, with self-concept showing more direct effects than self-satisfaction. Eschbach suggested that this finding was due to the global nature of the overall self-concept measure. Her findings also indicated that self-concept was a better predictor of IP than achievement tendency, effort, luck, ability, sex-role or career commitment.
Low self-esteem (or global self-worth) has repeatedly been associated with high IP scores. Kolligan (1990) suggested that imposters were unable to internalize positive events and had pessimistic, self-critical tendencies. By contrast, Harter (1989) contends that self-esteem is analogous to the construct of global self-worth which she defines as “the overall value that one places on the self as a person” (p. 67). She posits that self-esteem (global self-worth) is a merged self-assessment of (1) the discrepancy between a one's perceived success and one's ideal success and (2) the internalized perceptions of others’ approval. Brown (1993) described self-esteem as a strong generalized liking and positive regard for the self—”affection for the self” (p. 52).
Prince (1989) asserted that imposters are individuals who have strong emotional needs for validation and social support but are unable to get those needs met because of fears regarding making connections with others. He suggested that imposters often take a self-protective stance, distancing themselves from others. Cromwell's (1990) study with high school high achievers supported this picture of imposters.
It has been hypothesized that IP feelings begin sometime in childhood, yet there has not been research to support this. Clance and O’Toole (1987) suggest that false and non-affirming family messages contribute to the development of imposter feelings. Grays (1992) found retrospectively that adult imposters often felt misunderstood by emotionally distant parents. In view of the fact that the identification of IP feelings requires certain cognitive developmental advances (the ability to be introspective), research has primarily been limited to adults. However, adolescence is a developmental period when much self-reflection takes place. True self-awareness and introspection are poignant during adolescence and play an important role in identity formation (Erikson, 1950). Therefore, the study of IP during this developmental period is valuable in our understanding of the concepts and etiology of IP.
To date, only two studies have examined IP among adolescents. Stahl, Turner, Wheeler, and Elbert (1980) examined IP among high achieving African American high school girls and found that 55% attributed their achievements to reasons other than ability or intellect. Cromwell (1990) found that mean imposter scores among high achieving high school students were similar to previous mean scores using young adult samples. Based on these studies it appears that imposter feelings are well established by adolescence. The purpose of this study was to extend previous IP adolescent research and to ascertain the best predictors of high IP scores based on the variables of gender, global self-worth, social support, and self-concept.
Method
Participants
The participants for this study were 136 high school juniors and seniors (11th and 12th grades) from two suburban high school in Tulsa, OK. They were recruited from English, sociology and advisory classes. There were 71 females and 65 males; 4 African Americans, 2 Asians, 113 Caucasians, 2 Hispanics, 9 Native Americans, and 6 Others. Information regarding participants who declined to participate was not available. Informed consent forms were obtained from all students at the time of the study; those students under 18 years of age were also required to obtain a parent's or guardian's informed consent prior to the study. School permission was granted for both high schools several months prior to the study.
Measures
All participants were asked to complete a Demographic Data Form and four questionnaires. The questionnaires were the Harvey Imposter Phenomenon Scale, the Self-Perception Profile for Adolescents, the Multidimensional Test of Self-Concept, and the Social Support Scale for Children and Adolescents.
Demographic data form
A demographic data form was designed by the researchers to collect information about subjects’ gender, grade level, race, and perceived grade point average (GPA).
Harvey IP scale
The Harvey IP scale is a 14-item self-report instrument developed by Harvey (1981) to validate IP as a psychological construct. Each item is ranked on a 7-point Likert-type scale ranging from “not at all true” to “very true.” Examples of items are “In general, people tend to believe that I am more competent than I really am” and “I feel I deserve whatever honors, recognition, or praise I receive.” Total IP scores range from 0 to 84.
Harvey (1981) found an average inter-item reliability of .85 for 74 graduate students. Topping and Kimmel (1985) found an internal consistency of .75 using Cronbach alphas with a sample of university faculty members. No test–retest reliability information is currently available.
The self-perception profile for adolescents
The self-perception profile for adolescents is a 45-item self-report inventory developed by Harter (1988). The self-perception profile for adolescents measures eight specific domains of self-perception and a global self-worth dimension. Predictive, convergent, discriminant and construct validity have been demonstrated with this instrument. Using four different samples Harter (1988) found that eight of the nine subscales define their own factors. Global Self-Worth was not included in the factor analysis because, although it is qualitatively different from domain-specific perceptions, it is partially determined by how competent one feels in specific, highly valued domains.
It has been found to be a valuable instrument for purposes of theory development, for program evaluations and for individual clinical purposes. For purposes of this study, the domain used was Global Self-Worth since a more overall index of self-perception was the focus of interest for this study.
The social support scale for children and adolescents
The social support scale for children and adolescents is a 24-item self-report questionnaire developed by Harter (1985). It measures four possible sources of social support: parents, teachers, classmates and close friends. Internal consistency reliabilities for each subscale were initially established with two samples of elementary students and two samples of middle school students. For each subscale reliability ranges were: Parent .78–.88, Teacher .81–.84, Classmate .74–.79, Close Friend .72–.83 (Harter, 1985). Several questions on the Classmate and Close Friend subscales were rewritten in order to improve reliabilities. Factor analyses revealed four significant factors from the middle school students. No test–retest information is currently available.
The multidimensional test of self-concept (MTS)
The multidimensional test of self-concept is an 18-item self-report instrument developed by Lathrop in 1987. It measures one's current state (As You Really Are) and one's desired state (As You Would Like to Be) in the self-concept areas of Sociability (close-distant), Competence (inexpert–expert) and Dependability (reliable–unreliable) (Lathrop, 1987). Only the “As you Really Are” questionnaire was used for this study. Reliability was investigated using subjects ranging from 14 to 73 years of age (Lathrop, 1987). Generalizability coefficients for each of the subscales were: Sociability .86, Competence .78 and Dependability .83. Internal consistency coefficients (Alphas) for each subscale were: Sociability .84, Competence .72 and Dependability .79. Test–retest analysis after 3 months ranged from .66 to .71 for the “Are” scores.
Construct validity has been supported through factor analysis, confirming the three factors of Sociability, Competence and Dependability. Intercorrelations ranged from .41 to .54 (Lathrop, 1988). Path analysis was also used with all three “As You Really Are” subscales in determining precursers to academic self-esteem (Hudson, 1988).
Procedure
Following institutional approval, three teachers allowed the researchers to distribute questionnaires among their junior and senior level classes. Seven classes were introduced to the research project. After obtaining informed consent from the participants and/or their parents, participants completed the questionnaires. Incentives of suckers and an entry for a free movie pass were offered. Questionnaires were returned anonymously to the researchers.
Results
Gender comparisons
The total sample had a mean score of 31.15 with a standard deviation of 11.03 and a range of 8–69 on the Harvey IP Scale. The IP scores ranged from 8 to 69.
Females in this study had a mean score of 31.21 with a standard deviation of 10.95 while males had a mean score of 31.09 with a standard deviation of 11.21. However, these differences were not significant (
; 
).
Descriptive statistics and correlations
Correlations were run between all the variables and IP scores for all subjects and then for females and males separately. As seen in Table 1 IP scores were significantly negatively related to all of the variables (Global Self-Worth, Parent Support, Classmate Support, Teacher Support, Friend Support, Sociability, Competence, and Dependability) for both the total sample and for females. Correlations ranged between −.36 and −.50 among the total sample and between −.53 and −.63 for females. For males Parent Support and Teacher Support did not reach significance as a negative correlate. Correlations ranged between −.22 and −.44 for males.
| Variable | Pearson r correlations with imposter score | Comparison of z scores by gender | ||
|---|---|---|---|---|
Total ( ) |
Females ( ) |
Males ( ) |
||
| Global Self-Worth | −.47** | −.55** | −.39** | −1.09 |
| Parent Support | −.39** | −.54** | −.22 | −2.16* |
| Classmate Support | −.50** | −.63** | −.34** | −2.20* |
| Teacher Support | −.36** | −.53** | −.23 | −2.03* |
| Friend Support | −.50** | −.58** | −.44** | −1.08 |
| Sociability | −.49** | −.58** | −.38** | −1.49 |
| Competence | −.46** | −.54** | −.38** | −1.16 |
| Dependability | −.41** | −.54** | −.26* | −1.92 |
- *p<.05; **p<.01.
Pearson r's were transformed to Fisher's z’; two-tailed z-test comparisons were performed to test the significance of the difference between female and male IP correlations (see Table 1). Significant differences were found between male and female IP correlational scores with Parent Support, Classmate Support and Teacher Support (see Table 1).
Regression analyses
Three regression analyses were performed using IP as the dependent variable and Global Self-Worth, Parent Support, Classmate Support, Teacher Support, Close Friend Support, Sociability, Competence, and Dependability as independent variables. The first regression analysis used the entire sample while the second regression analysis examined only females and the third regression analyses examined only males.
Using the entire sample 40% of the variance in IP scores was accounted for by the variables (R2=.437; adjusted R2=.401). This was a statistically significant finding with 
(p<.001). Among the variables, Friendship Support (β=−.279; 
; p<.01) and Competence (β=−.181; 
; p<.05) were found to be significant predictors of IP scores. These results are shown in Table 2.
| Predictors | Standardized β | t | UVE |
|---|---|---|---|
| Global Self-Worth | −.078 | −.811 | .003 |
| Parent Support | −.057 | −.659 | .002 |
| Classmate Support | −.048 | −.561 | .006 |
| Teacher Support | −.113 | −1.154 | .002 |
| Friend Support | −.279 | −3.516*** | .055 |
| Sociability | −.107 | −1.137 | .006 |
| Competence | −.181 | −2.056** | .017 |
| Dependability | −.075 | −.924 | .004 |
| Overall F(8,127)=12.300** | R2=.437** | R2adj.=.401** | Shared variance=.342 |
- Note: UVE=Unique variance explained. *p<.10,**p<.05,***p<.01.
Also of interest in the total sample is the strength of the unique variance explained (UVE). While most UVEs were quite modest, Friend Support was moderate (.055) for the total sample. This is in contrast to Competence. While it was a significant predictor in the total regression model, it had a UVE of only .017.
For females, 56% of the variance in IP scores was accounted for the variables (R2=.613; adjusted R2=.563). This was a statistically significant finding with 
(p<.001). Among the variables, Friendship Support (β=−.217; 
; p<.05) and Dependability (β=−.228; 
; p<.05) were found to be significant predictors of IP scores. Classmate Support approached significance (β=−.260; 
; p<.06). Although these three variables were significant predictors in the overall model, the UVE for each was small. These results are shown in Table 3.
| Predictors | Standardized β | t | UVE |
|---|---|---|---|
| Global Self-Worth | −.058 | .458 | .001 |
| Parent Support | −.131 | −1.224 | .010 |
| Classmate Support | −.260 | −1.938* | .024 |
| Teacher Support | −.074 | −.678 | .003 |
| Friend Support | −.217 | −2.035** | .026 |
| Sociability | −.098 | −.839 | .005 |
| Competence | −.121 | −.967 | .006 |
| Dependability | −.228 | −2.262** | .032 |
| Overall F(8,62)=12.268*** | R2=.613*** | Total R2adj.=.563*** | Shared variance=.506 |
- Note: UVE=Unique variance explained. *p<.10,**p<.05,***p<.01.
For males, 23% of the variance in IP scores was accounted for by the variables (R2=.324; adjusted R2=.227). This was a statistically significant finding with 
(p<.01), Among the variables, only Friendship Support (β=−.326; 
; p<.05) was found to be a significant predictor of IP scores. Friendship Support also had a moderate UVE value (.082). These results are shown in Table 4.
| Predictors | Standardized β | t | UVE |
|---|---|---|---|
| Global Self-Worth | −.207 | −1.406 | .024 |
| Parent Support | .050 | .339 | .002 |
| Classmate Support | .040 | .256 | .001 |
| Teacher Support | −.088 | −.661 | .006 |
| Friend Support | −.326 | −2.597** | .082 |
| Sociability | −.139 | −.901 | .010 |
| Competence | .161 | −1.177 | .017 |
| Dependability | .040 | .293 | .001 |
| Overall F(8,56)=3.348*** | R2=.324*** | Total R2adj.=.227*** | Shared variance=.181 |
- Note: UVE=Unique variance explained. *p<.10,**p<.05,***p<.01.
Discussion
Mean scores on the Harvey IP Scale support previous research (Stahl et al., 1980; Cromwell, 1990) that IP exists among adolescents at about the same rate as in adult populations. This finding provides evidence that IP feelings are already well established by adolescence and that there may be earlier etiology of IP. This substantiates Clance and Ime's (1978) and Clance and O’Toole's (1987) original hypothesis that IP begins early in childhood and may have to do with family of origin dynamics.
Descriptively, the mean scores on the IP scale also suggest that the experience of having imposter thoughts and feelings is not gender-specific. This supports the majority of the research studies regarding gender and IP, but is contrary to Clance and Imes’ (1978) original hypothesis that girls experience feelings of fraudulence in greater proportion than boys. It would appear that adolescent boys as well as girls experience feelings of fraudulence at approximately the same rate.
Regression analyses examining the predictability of social support for IP scores by gender indicated that both adolescent females and males appear to need the support of close friends in order to avoid imposter feelings. When normal imposter feelings arise from new or challenging tasks, if one has a close friend with whom to discuss these and discover that the other, too, has faced such thoughts and feelings, a sense of connection and authenticity takes place. Miller and Stiver (1997) suggest that when someone “goes with us in the feelings” we are more able to accept our feelings as legitimate (p. 33). This experience of genuinely expressing thoughts and feelings has the effect then of moving one into a place of greater confidence and assurance.
Having a close friend may be particularly poignant for the adolescent in order to avoid imposter feelings in that peer relationships are generally becoming increasingly important (Buhrmester & Furman, 1987), at the same time that identity development is taking place (Erikson, 1950). Friendships during adolescence are recognized as a most important source of support as self-disclosure reaches its highest point during this developmental stage (Parker & Gottman, 1989; Furman & Buhrmester, 1992). Certainly these are complex processes taking place during adolescence where self-definition and interpersonal relatedness develop side by side.
Despite the gender similarities, a gender-specific picture does emerge for adolescent imposters. Adolescent female and male imposters appear to have differing self-perceptions and differing social support deficits. Factors which may lead adolescent females to be more vulnerable to IP may not be the same factors that lead adolescent males to experience IP. Females’ imposter feelings appear to be more inclusive of multiple support systems. While the lack of Friend Support is clearly associated with IP for females and males alike, for females it is only one predictor among many operating in similar ways. Based on comparisons of the male and female regression analyses for all variables and z score comparisons, adolescent females also appear to need the support of parents, teachers and classmates to buffer them against IP, although the adult support (parent and teacher) appears to be mediated by some other variable(s). It appears that adolescent girls need to connect with peers both generally (Classmate Support) and intimately (Friendship Support) in order to feel genuine about themselves and avoid feeling like imposters.
Regression analyses of specific measures of self-concept indicate that adolescent females also seem to need a sense of personal dependability in order to develop feelings of authenticity rather than fraudulence. This finding suggests that the female adolescent's need for a sense of connection in order to avoid IP may be broader than the male adolescent's. Dependability implies a relationship with another. It suggests that others find the self to be trustworthy, reliable and steadfast. These findings, along with findings that there appears to be shared variance among the self-concept and social support variables for females, suggest that females may have greater interdependent relationships and self-concepts. This supports feminist theory which suggests that women's healthy development happens in interdependent relationships where both connection and empowerment take place (Miller & Stiver, 1997). If adolescent females do not perceive themselves as personally dependable and able to develop relationships with multiple others they are more likely to have higher IP scores.
Adolescent males, by contrast, appear to depend more exclusively on Friend Support in order to feel authentic themselves. Findings that the UVE for Friend Support approached moderate strength, suggests that Friend Support may provide a more differentiated role in predicting IP for males. Their development of an authentic sense of self does not seem to be as connected to the support of adults or classmates in the same way that it does for females.
Additionally, social and self-concept factors seemed to contribute less to IP for adolescent males than for adolescent females. The predictor set used in this study (self-worth, social support, and self-concept) account for less than half of the amount of the variance in IP scores that the same predictor set accounted for in female IP scores. What can be accounting for higher IP scores in some males? Past research has suggested that physical and mathematical performance and higher achievement are increasingly important for adolescent males, but self-esteem does not show the after age nine drop in males that it does in females (Ruble & Martin, 1998). Future research should explore the predictability of these variables to measures of IP.
Among social support variables, a surprise finding was that parent support was not a significant predictor of IP scores. Previous research has found parent support to make a significant contribution to adolescents’ global self-esteem (Harter, 1989) and to true/false self-behavior (Harter, Marold, Whitesell, & Cobbs, 1996). While differing constructs and differing instrumentation may account for these seemingly contradictory findings, it may be that while parent support continues to play an important part in adolescents’ behavior and sense of value, it plays less of a part in adolescents’ developing sense of authenticity and identity.
A limitation of this study is, of course, the correlational nature of the identified associations. Correlational relationships do not provide information regarding causation. It could just as easily be argued that imposter feelings contribute to the distancing of oneself in social relationships rather than poor social relationships contributing to imposter feelings. Further research is needed to examine these relationships.
Since the IP is hypothesized to begin in childhood, it is suggested that future studies be done to determine the incidence of IP at younger ages. Investigations which examine earlier childhood experiences of IP as it pertains to self-perceptions and social support would also add to the field of knowledge. Further research is also needed in order to determine if poor peer connectedness is a developmental aspect of IP or one which is evident throughout the life span.
