Preliminary Results of a Scale Assessing Instability of Self-esteem
The objective of the study was to examine the psychometric properties of a new self-report questionnaire, the Instability of Self-Esteem Scale (ISES), designed to be a complement to the Rosenberg Self-Esteem Scale (RSES), which measures global self-esteem. A total of 383 nonclinical adolescents participated in an assessment consisting of the ISES, the RSES, and two measures of psychopathology, the Center for Epidemiological Studies-Depression scale (CES-D), and the Borderline Personality Inventory (BPI). The ISES had satisfactory internal consistency and test-retest stability. The ISES and RSES were weakly correlated, this finding suggesting that they measure quite independent constructs. The different pattern of intercorrelations of the ISES and RSES with the CES-D and BPI provides support for the divergent validity of the ISES.
La présente étude avait pour but d’examiner les propriétés psychométriques d’un nouveau questionnaire d’autoévaluation, l’Instability of Self-Esteem Scale (ISES), qui a été conçu pour devenir un complément de l’échelle de l’estime de soi de Rosenberg [Rosenberg Self-Esteem Scale (RSES)], qui mesure l’estime de soi globale. Trois cent quatre-vingt-trois adolescents issus de la population non clinique ont participé à une évaluation composée de l’ISES, de la RSES et de deux mesures de la psychopathologie, la Center for Epidemiological Studies-Depression scale (CES-D) et le Borderline Personality Inventory (BPI). L’ISES a démontré une cohérence interne et une stabilité test-retest satisfaisantes. La corrélation entre l’ISES et la RSES étant faible, nous émettons comme hypothèse que les deux échelles mesurent des construits qui sont indépendants les uns des autres. Les différences entre le modèle des intercorrélations de l’ISES et de la RSES par rapport au modèle des intercorrélations de la CES-D et du BPI viennent appuyer la validité divergente de l’ISES.
Stability of self-esteem has been shown to be an important component of self-esteem (Kernis, Cornell, Sun, Berry, & Harlow, 1993) and instability of self-esteem should be considered an important psychological variable (de Man & Becerril Gutiérrez, 2002). Instability of self-esteem is a risk factor for depression, particularly for people experiencing life stress, (Kernis et al., 1998; Roberts & Kassel, 1997) and for suicidal ideation for people with low self-esteem (de Man & Becerril Gutiérrez, 2002). Instability in self-esteem has been linked to excessive concern regarding feelings about one’s self as well as higher sensitivity to evaluation originating from different events and/or individuals in one’s entourage (Crocker, Karpinski, Quinn, & Chase, 2003; Gable & Nezlek, 1998; Kernis et al.; Rosenberg, 1986; Seery, Blascovich, Weisbuch, & Vick, 2004).
All these studies on the instability of self-esteem have used similar methodology. Level of self-esteem was assessed with the Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965), a well-validated measure of global self-esteem in adults and adolescents (Bagley, Bolitho, & Bertrand, 1997; Winters, Myers, & Proud, 2002). Instructions were modified to ask how respondents felt at that moment, thus capturing state self-esteem. Participants completed the RSES several times (e.g., every day for six weeks). For each participant, self-esteem (in)stability was computed as the standard deviation of the respondent’s scores across the repeated measurements, reflecting the fluctuation in state self-esteem.
Currently, to our knowledge, there are two specific scales assessing directly instability of self-esteem. Rosenberg (1965) has elaborated a 5-item scale designed to assess perceived stability of self-esteem (e.g., “Does your opinion of yourself tend to change a good deal or does it always continue to remain the same?”). Kernis, Grannemann, and Barclay (1992) have developed a specific scale for the direct assessment of instability of self-esteem. Self-rated stability of self-esteem was assessed by having participants rate how much they thought they would change their (dis)agreement on a day-to-day basis with each of the items on Rosenberg’s Self-Esteem Scale. The scale range of answers was from 1 (very slight or no change at all) to 5 (extreme amount of change). The correlation between self-ratings of stability and “statistical stability” as calculated from multiple assessments of state self-esteem was weak (r = 0.22). The authors also used the Rosenberg’s Stability Scale. The correlation between this scale and “statistical” stability was even weaker (r = 0.13). To our knowledge, this is the only study comparing this scale with “statistical” stability, which should be regarded as the “gold standard” against which other measures should be compared. The authors’ interpretation is that people may not be very aware of how stable or unstable their self-esteem is. A limitation of these two scales may be indeed that they assess changes in selfesteem by directly asking the respondent his or her perception of changes in the level of state self-esteem. This may be a quite complex cognitive task. The cognitive complexity of these two scales is heightened by the complexity of the instructions for Kernis et al.’s (1992) scale and by the relative complexity of the wording for the Rosenberg’s Stability Scale. Cognitive complexity reduces measurement quality and limits validity (Shumway, Sentell, Unick, & Bamberg, 2004).
Given the importance of stability of self-esteem as a psychological variable, it would be interesting to have a valid questionnaire specifically assessing instability of self-esteem for use in clinical research or practice. Both clinical and longitudinal epidemiological studies suggest the relevance of instability of self-esteem in adolescence. Psychoanalysts emphasized the frequency of self-esteem instability in adolescence, manifested by the oscillation between feelings of superiority and inferiority (Kernberg, 1984). The longitudinal study conducted by Alsaker and Olweus (1992) confirmed that instability of selfesteem is greater in early adolescence.
A first appraisal of the validity of the measure was obtained by assessing its relationships with the RSES and measures of other related constructs. In adolescence, loss of self-esteem has been shown to be linked to internalizing disorder, particularly to depressive disorders (Winters et al., 2002), whereas instability of self-esteem is considered to be linked to borderline and narcissistic personality organizations (Kernberg, 1984). Borderline personality organization is characterized by instability of identity, self-esteem, affectivity, behaviours, and relationships. Borderline personality symptoms are frequent in adolescents (Bernstein et al., 1993; Chabrol, Montovany, Chouicha, Callahan, & Mullet, 2001). We hypothesized that our measure of instability of self-esteem will be more highly correlated to borderline personality symptoms than to depressive symptoms, whereas the RSES will be more correlated to depressive symptoms than to borderline symptoms. Therefore, we planned to study the relationships between the instability of self-esteem measure, the RSES, and measures of depressive and borderline symptomatology.
What follows are the results of a preliminary study evaluating some aspects of the reliability and validity of a new self-report questionnaire assessing level of self-esteem instability. The studied sample consisted of high-school students.
Participants and Procedure
Data were obtained from four randomly selected high schools in Toulouse and Rodez, France. Seventeen classes were randomly drawn from these high schools providing a sample of 461 participants. Participants provided written informed consent before completing the questionnaires. Questionnaires were completed during class time in the presence of a master’s level psychology student who presented the study and collected the questionnaires.
Questionnaires were anonymous; no compensation was offered. Of the 461 students, 383 (83%) completed all the items of the questionnaires. Subjects included 187 males (mean age = 17 ± 1.3, range = 15-19) and 196 females, (mean age = 17 ± 1.2, range = 15-19). A subsample of 50 participants completed the scales assessing self-esteem and instability of self-esteem on two separate occasions (10 days apart). Another sample of 19 young adults (10 males, 9 females; mean age = 24.8 ±4.6) completed the scale assessing instability of self-esteem once and the self-esteem scale on five different occasions (once a week for five weeks).
Development of the Instability of Self-Esteem Scale (ISES)
Given that cognitive complexity may be the major limitation of previous scales assessing instability of self-esteem, we chose to emphasize brevity, simplicity, and clarity of wording, instructions, and self-evaluation task asked to the respondent. The ISES, consisting of four items (presented in Table 1), is designed to be a complement to the Rosenberg Self-Esteem Scale.
This brief format was chosen within the perspectives of clinical practice or for use in community-based studies to make the ISES easier to complete when presented along with the RSES. The four items are worded in the same format. The specific wording of the items is organized through the balance of opposed thoughts or feelings about self-worth (e.g., “Sometimes I feel worthless; at other times I feel that I am worthwhile”). The instructions were the same as the RSES (“Please circle the appropriate number for each statement depending on whether you strongly agree, agree, disagree, or strongly disagree with it”).
Participants indicated their agreement with items on the same 4-point scale as the RSES items (0 = disagree strongly, 3 = agree strongly). Total scale scores range from O to 12.
The Rosenberg Self-Esteem Scale is a 10-item self-report questionnaire (Rosenberg, 1965). The scale range of answers is from O (disagree strongly) to 3 (agree strongly). With the exception of four items, a higher score indicates higher self-esteem; the remaining four items are worded in the negative direction to assess negative self-image. (For these four items, the scores are reversed during scoring.) Total scale scores range from O to 30.
Depressive and borderline symptomatology was assessed using the French versions of the Center for Epidemiological Studies-Depression scale (CES-D; Radloff, 1977), and the Borderline Personality Inventory (BPI; Leichsenring, 1999).
The CES-D is composed of 20 items. The respondent is asked to score each item by noting the frequency with which he or she had experienced the symptom during the preceding week. Responses are made on a 4-point scale ranging from O (rarely or none of the time) to 3 (most or all of the time). Total scale scores range from O to 60. A cutoff of 24 has been suggested to detect adolescents with “moderate/severe” depressive symptoms (Rushton, Forcier, & Schectman, 2002). Cronbach’s α of the CES-D in this sample was .87.
The BPI is a 53-item self-report questionnaire that explores a broad range of phenomenological manifestations of borderline personality disorders symptoms including affectivity and identity disturbances, fear of closeness, interpersonal instability, self-mutilative/suicidal behaviour, impulsive behaviour, dissociative symptoms, and psychotic symptoms (Leichsenring, 1999). Items were rated on a Likert scale with seven levels of agreement ranging from “entirely false” to “entirely true.” Total scale scores range from 0 to 318. The reliability of the BPI has been shown in adolescents (Chabrol et al., 2004). Cronbach’s α of the BPI in this sample was .92.
Evaluation of the Reliability of the Instability Self-Esteem Scale
Reliability of the ISES was assessed by test-retest reliability and internal consistency and compared to those of the RSES. Test-retest reliability was measured by the correlation between the scores on the two administrations. Internal consistency was determined by the Cronbach α coefficient and the interitem correlation analyses. Coefficient α reflects the intercorrelation between items but also the number of items as it increases as the number of items increases. Thus, the mean interitem correlation (MIC) was also calculated by taking the mean of the correlation for each possible pair of criteria of item within each scale. The MIC, which is independent from the number of items of the scale, facilitated comparison between the RSES and the ISES.
Evaluation of Some Aspects of Validity of the ISES
The concurrent validity of the ISES was assessed by the relation between the ISES and self-esteem instability evaluated across the repeated measurements among the participants who completed RSES on five occasions. For each participant, the standard deviation of RSES scores over these five assessments was calculated, reflecting the degree of fluctuation in state self-esteem, with higher standard deviation indicating greater instability. The correlation between ISES scores and standard deviation of RSES scores was calculated using Pearson r coefficient. The strength of the underlying relationships was measured with effect sizes. A Pearson correlation coefficient of .10 is considered to be a small effect, .30 a medium effect, and .50 a large effect (Cohen, 1992).
Criterion validity was assessed by the comparison of scores between boys and girls and between depressed and nondepressed participants to test the capacity of the ISES to discriminate these groups. Student t-test was used.
The convergent validity of the ISES was assessed through the intercorrelations between the ISES, the RSES, the CES-D and the BPI, calculated using Pearson r. The relative contribution of ISES and RSES scores to CES-D scores were studied by a multiple regression analysis.
Reliability of the ISES
Internal consistency. The Cronbach α for the ISES and the RSES were .89 and .76, respectively. The MIC values for the ISES and the RSES were .45 and .35, respectively.
Test-retest reliability. The Pearson correlation coefficients across the two administrations were .89 (p
Validity of the ISES
Concurrent validity. The Pearson correlation coefficient calculated between ISES scores and standard deviations of RSES scores over repeated measurements showed that the two measures of instability of self-esteem were highly correlated (r = 0.81, p
Gender effects. Table 2 shows the comparison of the ISES, RSES, CES-D, and BPI scores between boys and girls. The ISES, CES-D, and BPI scores were significantly higher in girls whereas the RSES score was significantly lower in girls.
Comparisons of the ISES and RSES scores in depressed and nondepressed participants. Table 3 presents the comparisons of the ISES and RSES scores in participants with and without moderate/severe depressive symptomatology. Among participants with moderate/severe depressive symptomatology, ISES scores were significantly higher whereas RSES scores were significantly lower.
Intercorrelations of the JSES, RSES, CES-D, and BPL All correlations were significant. The level of self-esteem and self-esteem instability were weakly negatively correlated (r = -20, p
Multiple regression analysis predicting CES-D scores. In a multiple regression analysis predicting CES-D scores in the total sample, we entered age, sex, ISES and RSES scores. This model accounted for 36% of the variance of CES-D scores, F(4, 378) = 53, p
This study describes a new scale assessing instability of self-esteem, which was designed as a complement to the RSES and reports preliminary data referring to reliability and validity.
The results suggest satisfactory reliability of the ISES. Internal consistency of the ISES was high and test-retest stability within a 10-day retest interval was satisfactory.
The concurrent validity was high, as measured by the correlation between the ISES and the variability of RSES scores across repeated measurements (r = .81). The ISES appeared to perform much better than the Self-rated stability of self-esteem scale (Kernis, et al., 1992) as well as with the RSES, the correlation of which with “statistical” stability was weak (r = .22 and r= .13, respectively).
The findings also provide some support for the criterion validity of the ISES as the ISES was capable of discriminating between boys and girls and between depressed and nondepressed subjects as is the RSES.
Girls had significantly lower self-esteem than boys. This result is in agreement with previous studies that have consistently shown that boys report higher self-concept than girls (e.g., Bagley et al., 1997; McDonald & McKinney, 1994). Rather than indicating “poorer” self-esteem among girls, this may reflect a feminine response style of being less self-declaratory on the “powerful” aspects of self (Bagley et al., 1997).
In the present study, girls also had significantly higher instability of self-esteem than boys. This is consistent with the findings of the longitudinal study conducted by Alsaker and Olweus (1992). These results may be also accounted for by girls’ response style. These findings may also be linked to the higher depressive and borderline symptomatology reported by girls, a result which is also in keeping with previous studies (e.g., Chabrol et al., 2004; Rushton et al., 2002). Given the high correlation between depressive and borderline personality symptomatology and instability of self-esteem, the higher instability in girls may reflect their higher level [propensity for] of depressive and borderline symptoms.
The comparisons of RSES and ISES scores between depressed and nondepressed participants suggest that self-esteem is both lower and more unstable in depressed participants. This result compares with the studies that found an association between instability of self-esteem and depression (Kernis et al., 1998; Roberts & Kassel, 1997).
The results of the regression analysis suggest that levels of self-esteem and instability of self-esteem contributed independently to the level of depressive symptomatology. However, no causality can be inferred from these data. A possibility is that low level of self-esteem and high level of instability may be consequences of high depressive symptomatology.
The degrees of correlation between the four studied scales argue for the construct and convergent validity of the ISES. The weak correlation between the ISES and RSES suggests that they measure different constructs. The different patterns of correlation of the ISES and RSES with the CES-D and BPI provide support for the divergent validity of the ISES. The level of instability of self-esteem was more associated with the level of borderline symptomatology than with the level of depressive symptomatology, whereas the level of self-esteem was more associated with the level of depressive symptomatology than with the level of borderline symptomatology.
Level of global self-esteem as measured by the RSES and instability of self-esteem as measured by the ISES appeared to be related but quite independent dimensions of self-esteem in adolescence.
The ISES may be a helpful instrument to explore an important dimension of self-esteem. Using both RSES and ISES might evaluate self-esteem more fully for adolescents in both psychological research and clinical assessment without the need for repeated measures of self-esteem alone. The next step in the validation of the ISES would be to use both the ISES and statistical stability to replicate findings from previous research that used only statistical stability.
Alsaker, R D., & Olweus, D. (1992). Stability of global self-evaluation in early adolescence: A cohort longitudinal study. Journal of Adolescence, 2, 123-145.
Bagley, C., Bolitho, E, & Bertrand, L. (1997). Norms and construct validity of the Rosenberg Self-Esteem Scale in Canadian high school populations: Implications for counselling. Canadian Journal of Counselling, 31, 82-92.
Bernstein, D. P., Cohen, P., Velez, C. N., Schawb-Stone, M., Siever, L. J., & Shinsato, L. (1993). Prevalence and stability of the DSM-III-R personality disorders in a community-based survey of adolescents. American Journal of Psychiatry, 150, 1237-1243.
Chabrol, H., Montovany, A., Chouicha, K., Callahan, S., & Mullet, E. (2001). Frequency of borderline personality disorder in a sample of French high school students. Canadian Journal of Psychiatry, 46, 847-849.
Chabrol, H., Montovany, A., Chouicha, K., Duconge, E., Kallmeyer, A. Mullet, E., et al. (2004). Factor analyses of the Borderline Personality Inventory in adolescents. European Journal of Psychological Assessment, 20, 59-65.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
Crocker, J., Karpinski, A., Quinn, D. M., & Chase, S. K. (2003). When grades determine self-worth consequences of contingent self-worth for male and female engineering and psychology majors. Journal of Personality and Social Psychology, 85, 507-516.
de Man, A. E, & Becerril Gutiérrez, B. I. (2002). The relationship between level of self-esteem and suicidal ideation with stability of self-esteem as moderator. Canadian Journal of Behavioural Science, 34, 235-238.
Gable, S. L., & Nezlek, J. R. (1998). Level and instability of day-to-day psychological well-being and risk for depression. Journal of Personality and Social Psychology, 74, 129-138.
Kernberg, O. (1984). Severe personality disorders: Psychotherapeutic strategies. New Haven, CT: Yale University.
Kernis, M. H., Cornell, D. P., Sun, C. R., Berry, A., & Harlow, T. (1993). There’s more to self-esteem than whether it is high or low: The importance of stability of self-esteem. Journal of Personality and Social Psychology, 65, 1190-1204.
Kernis, M. H., Grannemann, B.D., & Barclay, L. (1992). Stability of self-esteem: Assessment, correlates, and excuse making. Journal of Personality, 60, 621-644.
Kernis, M. H., Whisenhunt, C. R., Waschull, S. B., Greenier, K. D., Berry, A.J., Herlocker, C. E., et al. (1998). Multiple facets of self-esteem and their relations to depressive symptoms. Personality and Social Psychology Bulletin, 24, 657-668.
Leichsenring, F. (1999). Development and first results of the Borderline Personality Inventory: A self-report instrument for assessing borderline personality organisation. Journal of Personality Assessment, 73, 45-63.
McDonald, D., & McKinney, P. (1994). Steady dating and self-esteem in high school students. Journal of Adolescence, 17, 557-564.
Radloff, L. S. (1977). The CES-D scale: A self report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401.
Roberts, J. E., & Kassel, J. D. (1997). Labile self-esteem, life stress, and depressive symptoms: Prospective data testing a model of vulnerability. Cognitive Therapy and Research, 21, 569-589.
Rosenberg, M. (1965). Society and adolescent self-image. Princeton, NJ: Princeton University Press.
Rosenberg, M. (1986). Self-concept from middle childhood through adolescence. In J. Suls & A. G. Greenwald (Eds.), Psychological perspectives on the self (Vol. 3, pp. 107-135). Hillsdale, NJ: Erlbaum.
Rushton, J. L., Forcier, M., & Schectman, R. M. (2002). Epidemiology of depressive symptoms in the National Longitudinal Study of Adolescent Health. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 199-205.
Seery, M. O., Blascovich, J., Weisbuch, M., & Vick, S. B. (2004). The relationship between self-esteem level, self-esteem stability, and cardiovascular reactions to performance feedback. Journal of Personality and Social Psychology, 87, 133-145.
Shumway, M., Sentell, T., Unick, G., & Bamberg, W. (2004). Cognitive complexity of self-administered depression measures. Journal of Affective Disorders, 83, 191-198.
Winters, N. C., Myers, K., & Proud, L. (2002). Ten-year review of rating scales. III: Scales assessing suicidality, cognitive style, and self-esteem. Journal of the American Academy of Child and Adolescent Psychiatry, 4, 1150-1181.
Received November 2, 2004
Revised June 16, 2005
Accepted July 14, 2005
HENRI CHABROL, AMÉlie ROUSSEAU, and STAGEY CALLAHAN, Université de Toulouse-Lé Mirail
Correspondence concerning this article should be addressed to H Chabrol, 21 rue des Cèdres, 31400 Toulouse, France (Tel.: 33 561 22 52 90; Fax: 33 561 25 70 93; E-mail: email@example.com).
Copyright Canadian Psychological Association Apr 2006
Provided by ProQuest Information and Learning Company. All rights Reserved