The Rosenberg Self-Esteem Scale: Scoring, Normal Ranges, and Clinical Use

The Rosenberg Self-Esteem Scale was created by the sociologist Morris Rosenberg and published in 1965 in Society and the Adolescent Self-Image (Princeton University Press). Rosenberg developed it with a sample of more than 5,000 New York high school students, working on questions about adolescent identity, social class, and religious background rather than on clinical assessment.

Sixty years later it remains the most widely used measure of global self-esteem in research, appearing in thousands of studies across dozens of countries and in nearly every literature that touches self-concept: depression, eating disorders, body image, adolescent development, organizational psychology, and social comparison research. Few instruments in psychology have that reach, and almost none of them are ten items long.

Part of the reason for its persistence is practical. The scale takes about two minutes, requires no training to administer, and is free to use with attribution to the Rosenberg estate. Part of it is that Rosenberg's construct held up. Global self-esteem, defined as an overall favorable or unfavorable attitude toward the self, turned out to be a stable and measurable thing that predicts outcomes decades out.

This site hosts a free Rosenberg Self-Esteem Scale that handles the reverse scoring and returns a banded result, which removes the most common source of hand-scoring error.

Internal consistency typically falls between .77 and .88 across samples, which is strong for a ten-item unidimensional measure. Test-retest reliability sits around .82 to .85 over short intervals, consistent with the trait-like stability Rosenberg described.

That stability matters for how the scale is read. A measure with test-retest reliability in the mid .80s over a week or two is telling you about a durable attitude, so a five-point swing between two administrations a fortnight apart is more likely to reflect measurement noise or a bad day than a real change in self-regard.

Factor structure has been argued about for decades. The scale is intended as unidimensional, but the positively and negatively worded items often load on separate factors in confirmatory analyses, an artifact of item wording that has generated a small literature of its own. For clinical purposes the total score is the number to use.

The scale has ten items, half positively worded and half negatively worded, each answered on a four-point agreement scale running from Strongly Agree to Strongly Disagree.

In the common 0-3 scoring, positively worded items are scored Strongly Agree 3, Agree 2, Disagree 1, Strongly Disagree 0. The five negatively worded items (items 2, 5, 6, 8, and 9 in the standard ordering) are reverse scored, so agreement with a statement such as feeling useless at times earns 0 rather than 3. Summing all ten items gives a total from 0 to 30.

Conventional interpretation, which this site's tool displays: 0-14 Low Self-Esteem, 15-25 Average Self-Esteem (scores in this range are typical), 26-30 High Self-Esteem. Note how wide the average band is. Eleven of the thirty-one possible scores fall inside it, and most people land there.

Alternative scoring systems exist in the literature, including a 1-4 version producing a 10-40 range. The bands above apply to the 0-30 version only. Before comparing a client's score to a published figure, check which scoring the source used, because a 22 means different things on the two scales.

The Rosenberg is a descriptive measure of a psychological attitude. It screens for nothing, diagnoses nothing, and has no clinical cutoff validated against any disorder. A low score is a clinical signal worth exploring in session, and that is the whole of its diagnostic weight.

Global self-esteem is a trait-like attitude toward the self as a whole, distinct from the domain-specific self-evaluations people hold about their competence at work, their appearance, or their worth as a parent. Someone can rate themselves highly as a professional and still return a low Rosenberg score, because the scale asks about the person underneath the roles.

Self-esteem and depression are related without being the same construct. They correlate substantially, low self-esteem prospectively predicts later depressive episodes, and a depressive episode reliably drags self-esteem scores down while it lasts. A low Rosenberg score in a currently depressed client tells you relatively little on its own, since the state is contaminating the trait measure. The same score in a client whose PHQ-9 sits at 3 is far more informative.

The balanced wording, five positive items and five negative, was an early defense against acquiescence bias, the tendency some respondents have to agree with whatever a questionnaire puts in front of them. A pure agreement bias cancels out across the two halves. The design predates most modern discussion of response styles by decades, and it is one of the reasons the instrument aged well.

Cultural response styles shift the means. Comparisons across countries have repeatedly found different average scores that track cultural norms about self-presentation and modesty as much as they track self-regard, so a raw score from one cultural context does not translate cleanly into another. Interpreting a recently arrived client's 13 against a US undergraduate mean will mislead you. Within a person, tracked over time, the score behaves far better than it does across groups.

No score on this scale diagnoses anything. There is no self-esteem disorder in the DSM. Low self-esteem shows up as a feature across depression, social anxiety, eating disorders, personality disorders, and the aftermath of chronic invalidation, and identifying which of those is generating it is interview work.

Screening and case formulation: administered at intake, the Rosenberg gives a quick read on how a client regards themselves, which frequently reframes a presenting complaint. A client who came in about work stress and returns a score of 9 is describing a different problem than the intake paperwork suggested.

Monitoring cadence: because self-esteem is trait-like and moves slowly, monthly or quarterly administration is enough during active treatment. Weekly administration mostly produces noise. Symptom measures such as the PHQ-9 handle the week-to-week tracking; the Rosenberg is there to answer a slower question about whether the underlying self-view has shifted.

Meaningful change: the scale has no established reliable change index, so read movement conservatively and in context. Three or four points inside the same band across two administrations is unremarkable. A move from 11 to 19 sustained over several months, crossing from the low band into average, is the kind of change worth naming in session and worth asking the client to account for in their own words.

Measurement-based care: self-esteem work is notoriously hard to see progress in, both for clients and for clinicians. A number recorded at intake and revisited at month three gives the conversation something concrete to stand on, particularly with clients whose self-report on good days contradicts their self-report on bad ones. The score becomes evidence in an argument the client has been losing to themselves.

Treatment targets follow from the item pattern more than from the total. Items about self-respect and worth point toward core-belief work and self-compassion practice, while items about capability and satisfaction with the self point toward behavioral experiments and mastery experiences.

0-14 (Low Self-Esteem): the client's overall attitude toward the self is unfavorable, and at the lower end of this band the endorsements are usually stark. Scores here co-occur with depression, social anxiety, eating disorders, and histories of chronic criticism or invalidation, so the first clinical question is what is generating the low regard rather than how to raise the number. Sustained low scores respond to core-belief work, self-compassion training, and mastery experiences, and they typically move over months.

15-25 (Average Self-Esteem): the range where most people score, including most people who are doing well and a fair number who are not. Rosenberg's construct is an ordinary human attitude, and ordinary means mixed: some pride, some doubt, a few items endorsed in each direction. Scores here rarely become a treatment target on their own. Movement within the band during treatment is usually noise.

26-30 (High Self-Esteem): a consistently favorable self-view, endorsed at or near the ceiling. High scores are generally a protective factor, associated with resilience and better outcomes under stress. Read the very top of the range with a little care in specific clinical contexts, since a perfect 30 in someone whose presentation is visibly strained can reflect impression management or defensive self-enhancement rather than settled self-regard.

Is a Rosenberg score of 15 bad? No. Fifteen sits at the bottom of the average band, one point above the low range, which means it describes a typical rather than a deficient level of self-regard. The score is close enough to the boundary that it is worth a follow-up question about which items drove it, and it does not by itself indicate a problem needing treatment.

What does a Rosenberg score of 12 mean? Twelve falls in the low band, roughly at the midpoint of it. It indicates that the person endorsed unfavorable views of themselves across several items over a general reference period. The number carries no diagnosis; it flags an attitude toward the self worth understanding, and the useful next step is reading which items were endorsed and asking about them directly.

What is a normal Rosenberg score? Anything from 15 to 25 counts as typical, and most of the general population falls in that span. Published means from Western student and community samples usually land somewhere around the low twenties on the 0-30 scoring, though those figures shift with country and sample, so treat them as a rough anchor.

Which items are reverse scored? Items 2, 5, 6, 8, and 9 in the standard ordering, the five negatively worded statements. Reversing them is the step most often missed in hand scoring, and missing it inverts the result: a client with genuinely high self-esteem produces a low total, and the clinician goes looking for a problem that the arithmetic invented.

Can self-esteem scores change with treatment? Yes, though slowly compared with symptom measures. A PHQ-9 can drop ten points in six weeks of behavioral activation. A Rosenberg score of 10 that reaches 18 usually took months of accumulated evidence against a long-held belief. Clients should hear that timeline at the outset, because expecting fast movement on this measure sets up a discouraging comparison with the symptom scores that are moving faster.

How does the Rosenberg compare with symptom scales like the PHQ-9? They measure different layers. The PHQ-9 counts depressive symptoms over the past two weeks and moves with episodes. The Rosenberg reads a durable attitude toward the self and moves with the underlying belief. Running both is informative precisely because they can diverge: symptoms remitting while self-esteem stays flat is a common pattern near the end of treatment, and it often marks the work that remains.

Is there a cutoff for clinically low self-esteem? No validated one. The 15 boundary between the low and average bands is a widely used convention, not a threshold established against any clinical criterion, and it should be read as descriptive. Treat scores near the boundary as continuous rather than categorical, and let the clinical interview settle whether the self-view is doing damage.

Rosenberg, M. (1965). Society and the Adolescent Self-Image. Princeton, NJ: Princeton University Press.

Rosenberg, M. (1979). Conceiving the Self. New York: Basic Books.

Sinclair, S. J., Blais, M. A., Gansler, D. A., Sandberg, E., Bistis, K., and LoCicero, A. (2010). Psychometric properties of the Rosenberg Self-Esteem Scale: overall and across demographic groups living within the United States. Evaluation and the Health Professions, 33(1), 56-80.

Schmitt, D. P., and Allik, J. (2005). Simultaneous administration of the Rosenberg Self-Esteem Scale in 53 nations. Journal of Personality and Social Psychology, 89(4), 623-642.