Self-Esteem Assessment

Self-esteem is one of the most widely studied variables in the social and behavioral sciences. A vast literature spanning many disciplines has shown that high self-esteem promotes behaviors, goals, and coping mechanisms that facilitate success in school, work, and relationships. High self-esteem impedes mental and physical health problems, delinquency, substance abuse, and antisocial behavior. Given its importance, it is not surprising that researchers have developed a diverse array of scales and procedures to assess self-esteem.

This article discusses the issues involved in assessing self-esteem and provides an overview of the leading self-esteem measures used by psychologists. The entry begins by explaining what psychologists mean by the construct of self-esteem. Next it provides an overview of the most commonly used measures of self-esteem. The entry ends by briefly addressing several broader questions that bear on the interpretation of self-esteem measures and their use in research and applied contexts, including, Is self-esteem a stable trait or a transient state? How large a concern is socially desirable responding? Can self-esteem be assessed in young children? Does self-esteem vary as a function of age, gender, social class, and ethnicity? How does self-esteem differ from other aspects of personality, such as narcissism and neuroticism?

Conceptualizing Self-Esteem

Self-esteem is generally conceptualized in terms of phenomenological experience. At the turn of the 20th century, William James defined self-esteem as the degree to which people perceive their accomplishments as consistent with their goals and aspirations. In the mid-1980s, Morris Rosenberg extended this definition by adding that self-esteem involves feelings of self-acceptance, self-liking, and self-respect. These definitions have been applied to both global and domain-specific self-esteem. Global self-esteem refers to an individual’s overall evaluation of him- or herself. Domain-specific self-evaluations focus on a specific facet of the self, such as physical appearance or academic competence.

The conceptualization of self-esteem as a phenomenological construct is inextricably linked to its measurement. If people subjectively experience themselves as competent, likable, and valued, then they necessarily have high self-esteem, regardless of their actual level of competence and likableness. Given this conceptualization, it is not surprising that self-esteem is most commonly assessed via self-report scales.

Commonly Used Measures of Self-Esteem

The measurement of self-esteem is a long-standing issue in psychology, dating back to James’s seminal writings on the self in 1890. Over the years, researchers have proposed a wide range of self-esteem measures. In addition to numerous self-report scales, there are “experience sampling” measures, pictorial and puppet measures for children, measures based on observer and peer ratings, self-ideal discrepancy measures, measures based on preferences for the letters in one’s name, and reaction time measures such as the Implicit Association Test (IAT). The vast majority of researchers rely on face-valid self-report scales (i.e., scales in which what is being assessed is obvious); therefore, this section focuses on several of the most widely used self-report scales. However, measures of implicit self-esteem are increasing in popularity, so this section also reviews the most commonly used implicit measure, the IAT. This section ends by discussing the degree to which different self-esteem scales and measures are correlated with each other (i.e., convergent validity).

Self-Report Scales

The most commonly used self-report scale is the Rosenberg Self-Esteem Scale, followed by the Coopersmith Self-Esteem Inventory, the Harter Self-Perception Profile, and the Marsh Self-Description Questionnaire. The Rosenberg scale assesses global self-esteem. The other three measures assess global and domain-specific self-evaluations, but in some research contexts, only the global self-esteem subscales are used.

Rosenberg Self-Esteem Scale (RSE).

In work with adults, the RSE is by far the most widely used measure of self-esteem, and it has received more psychometric analysis and empirical validation than any other self-esteem measure. The RSE is made up of 10 face-valid items, 5 positively worded and 5 negatively worded. Most researchers use a 4-, 5-, or 7-point Likert scale with the end points labeled strongly disagree to strongly agree. Several researchers have adapted or shortened the RSE. Most recently, a single-item version (SISE) was developed and validated for use when time is limited. The SISE asks the participant to rate the statement “I have high self-esteem” using a 5-point Likert scale.

Coopersmith Self-Esteem Inventory (SEI).

The SEI was developed to assess multiple self-concept domains in children. The original SEI assessed four domains (peers, parents, school, and personal interests) and included 50 items that were rated as being unlike me or like me. Coopersmith later created a modified version based on the best 25 items. The SEI is primarily used with children, but an adult scale is also available.

Self-Perception Profile (SPP).

The SPP was developed to assess multiple domains of self-concept and includes a separate global self-concept subscale. The original SPP was developed for children, but it has been adapted for use with adolescents, college students, and adults. Therefore, the SPP has the advantage of allowing cross-age comparisons.

The SPP presents two statements and asks the participant to choose the one that is most self-descriptive, and then rate it as a lot like me or a little like me. These responses are coded into a 4-point scale. The scale for young children (with preschool to first grade and boy and girl versions available) is administered individually in pictorial format, eliminating the need for reading skills. However, no global subscale is available for young children because it has been argued that children under 8 cannot communicate feelings of global self-worth. The other scales use the same format as the scale for young children, except pictures are not used. All scales assess multiple dimensions, but some of the dimensions vary to ensure that they are age appropriate.

Self-Descriptive Questionnaire (SDQ).

The SDQ was developed to assess domain-specific self-evaluations across age groups and includes a separate global sub-scale. The SDQ is longer than most other self-concept scales and it has separate forms for children (76 5-point Likert items), for adolescents (102 6-point Likert items), and for college students and adults (136 8-point Likert items). If cross-age comparisons are desired, these scales can be adapted to use the same response format. All scales assess multiple dimensions, but some of the dimensions vary to ensure they are age appropriate.

The Implicit Association Test (IAT)

The self-report scales described above may not tap true, deep-seated feelings about the self because they are face-valid. Therefore, participants might base their answers on what they believe the scale administrator wants to hear or what they believe is most socially desirable. To address this concern, researchers have developed measures of implicit self-esteem (i.e., aspects of self-evaluation that are assumed to be automatic, over-learned, and operative outside of conscious awareness). Such measures are believed to avoid the need for direct self-evaluation that occurs when individuals are asked to complete a self-report measure of self-esteem.

One of the most commonly used measures of implicit self-esteem, the IAT, uses a computerized reaction time procedure to assess the degree to which individuals associate the self with positive versus negative concepts. The IAT self-esteem score is based on the difference between the time it takes an individual to respond to positive items linked with the self versus negative items linked with the self. A person with high implicit self-esteem will link positive items with the self faster than negative items (i.e., they have an implicit representation of the self as positive). Because the task requires the individual to respond extremely quickly, the associations assessed through this procedure are assumed to be automatic and unconscious. If the IAT is administered along with a self-report measure of self-esteem, the IAT should be administered first to avoid carryover effects.

Correlations Between Measures

Different self-report measures of self-esteem tend to be moderately to strongly correlated with each other (rs = .40 to .70). Substantially lower correlations are found between self-esteem measures based on different assessment methods. Correlations between self and observer/informant reports of self-esteem generally range from the .20s to the .40s. Correlations between self-report scales and implicit measures of self-esteem are generally weak to nonexistent; that is, an individual’s self-reported level of self-esteem has virtually no relation to measurements of his or her implicit level of self-esteem. Researchers are currently debating whether this lack of convergence reflects problems with the validity of the implicit measures or the fact that explicit and implicit self-esteem are psychologically distinct constructs and therefore should not be empirically related. Findings such as this raise a number of broader issues regarding the use of self-esteem measures in research and real-world contexts. Some of these issues are addressed in the following section.

Broader Issues

Is Self-Esteem a Stable Trait or a Transient State?

Researchers who study the correlates and consequences of self-esteem generally assume it is a stable trait that predicts future behavior. The research literature provides ample support for this claim. Although self-esteem levels can be temporarily elevated or depressed in everyday life and through laboratory manipulations, numerous longitudinal studies have documented the stability of self-esteem across years and even decades of life. Test-retest correlations for self-esteem are comparable to those found for more basic dimensions of personality, such as extraversion, conscientiousness, and neuroticism. At the same time, the stability of self-esteem is far from perfect, even after taking into account the attenuating effects of measurement error (i.e., scale unreliability). Thus, self-esteem, like more basic personality traits, continues to change throughout the life span.

Is Socially Desirable Responding a Concern?

An important question is whether individuals with high self-esteem truly believe the highly desirable picture they paint of themselves, or if they are intentionally hiding their personality flaws and inflating their virtues. A related question is whether they are in fact responding honestly (i.e., basing their responses on conscious self-beliefs), but lack insight into their deep-seated, unconscious feelings. These two forms of social desirability—intentional self-presentation (or impression management) and self-deception—have been examined in relation to self-esteem.

Self-esteem is generally not associated with the impression management facet of social desirability, but it is associated with self-deception. Thus, high self-esteem individuals generally do not attempt to manipulate their responses to portray themselves in a positive light to others. However, they may be prone to self-deception because they are unaware of self-doubts about their competence, attractiveness, and likableness. Whether or not this poses a problem in the measurement of self-esteem depends on one’s conceptualization of self-esteem. If self-esteem is defined in terms of conscious experience, then the fact that conscious self-evaluations may not reflect unconscious feelings and beliefs does not pose a problem for the assessment of self-esteem. However, if unconscious feelings and beliefs are assumed to be central to self-esteem, then implicit measures might be more valid and useful than explicit measures.

Is Self-Esteem Distinct From Related Personality Constructs?

Self-esteem is empirically related to several basic dimensions of personality. High self-esteem individuals tend to score high on the traits Extraversion, Conscientiousness, Agreeableness, and Openness to Experience, and low on Neuroticism. Despite this empirical overlap, self-esteem often predicts important life outcomes over and above these basic personality dimensions.

A more complex issue is the relation between self-esteem and narcissism. Measures of self-esteem and measures of narcissism tend to be moderately correlated. However, theories of narcissism suggest that this overlap is not valid, but rather reflects a tendency for narcissistic individuals to have inflated self-esteem as a way of defending against unconscious feelings of inadequacy. From this perspective, self-esteem measures are contaminated by narcissism, because narcissistic individuals attain higher scores than is warranted by their true level of self-esteem. To correct for this problem, some researchers have conducted statistical analyses to empirically distinguish individuals whose self-esteem is inflated by narcissism from those who have genuinely high self-esteem. In many prediction contexts, the removal of narcissistic tendencies from measures of self-esteem actually increases the predictive validity of self-esteem and can even reverse the direction of self-esteem effects.

Assessing Self-Esteem In Young Children

Despite some divergence of opinion, most researchers believe that self-report, pictorial, or puppet-based measures of global self-esteem are valid for children as young as age 5 or 6, and virtually all researchers believe that global self-esteem scales are valid for children age 8 and older.

Differences In Self-Esteem as a Function of Age, Sex, Social Class, and Ethnicity

Self-esteem tends to be high in childhood, drops during adolescence (particularly for girls), increases gradually throughout adulthood, and then declines sharply in old age. Despite these changes in average levels, individuals tend to maintain their relative ordering: Individuals who have relatively high (or low) self-esteem at one point in time tend to have relatively high (or low) self-esteem years later. Nevertheless, this rank-order stability is somewhat lower during childhood and old age than during adulthood.

On average, men’s self-esteem scores tend to be about one third of a standard deviation higher than women’s self-esteem scores. This gender difference emerges during adolescence and holds across most of the life span. Socioeconomic status is very weakly related to self-esteem levels; wealthy individuals have only slightly higher self-esteem than working class individuals. In the United States, African Americans and Caucasians tend to report the highest levels of self-esteem, followed by Latinos/as and then Asian Americans.

A voluminous body of research suggests that self-report measures of self-esteem are useful tools in psychological assessment—they are reliable, valid in a wide range of populations, and predict important real-world outcomes. In short, if the goal is to discover whether individuals consciously experience positive self-regard, it may be possible, as the eminent psychologist Gordon Allport suggested, to simply ask them.

References:

Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. E. (2003). Does high self-esteem cause better performance, interpersonal success, happiness, or healthier lifestyles? Psychological Science in the Public Interest, 4, 1—14.
Blascovich, J., & Tomaka, J. (1991). Measures of self-esteem. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 115-160). San Diego, CA: Academic Press.
Byrne, B. M. (1996). Measuring self-concept across the lifespan. Washington, DC: American Psychological Association.
Harter, S. (1999). The construction of the self. New York: Guilford Press.
James, W. (1890). The principles of psychology. Cambridge, MA: Harvard University Press.
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg Self-Esteem scale. Personality and Social Psychology Bulletin, 27, 151-161.
Trzesniewski, K. H., Donnellan, M. B., Caspi, A., Moffitt, T. E., Robins, R. W., & Poultin, R. (2006). Adolescent low self-esteem is a risk factor for adult poor health, criminal behavior, and limited economic prospects. Developmental Psychology, 42, 381-390.