← All articles
Rosenberg Self-Esteem Scale Scoring and Interpretation

1 October 2026

Rosenberg Self-Esteem Scale Scoring and Interpretation

To score the Rosenberg Self-Esteem Scale, rate each of the 10 items from 0 to 3, reverse-score the five negatively worded items, and add them up. Totals run from 0 to 30 (or 10 to 40 if items are coded 1 to 4). Higher scores mean higher self-esteem. There is no official cut-off for low self-esteem.

That is the quick answer. Two things catch clinicians out. The item numbers you reverse depend on which copy of the scale you are holding, and the same client can score 12 or 22 depending on how the responses are coded. Both need to be settled before a total means anything.

What does the Rosenberg Self-Esteem Scale measure?

The Rosenberg Self-Esteem Scale (RSES) is a 10-item self-report measure of global self-esteem, published by the sociologist Morris Rosenberg in Society and the Adolescent Self-Image (1965). Five items are worded positively and five negatively, and each is answered on a four-point scale from strongly agree to strongly disagree.

According to the University of Maryland Department of Sociology, where Rosenberg was a professor, the scale was developed with 5,024 high school juniors and seniors from 10 randomly selected schools in New York State. The same page reports test-retest correlations typically between .82 and .88 and Cronbach's alpha between .77 and .88.

The instructions ask about general feelings about yourself, with no time frame such as the past two weeks. The RSES is therefore read as a fairly stable, trait-like measure, not a symptom snapshot.

How do you score the Rosenberg Self-Esteem Scale?

The University of Maryland's scoring instructions are:

  • Positively worded items: strongly agree = 3, agree = 2, disagree = 1, strongly disagree = 0
  • Negatively worded items: strongly agree = 0, agree = 1, disagree = 2, strongly disagree = 3
  • Total: add the 10 item scores, for a range of 0 to 30

One trap sits on the form itself. Some printed copies label the response columns 1 (strongly agree) to 4 (strongly disagree). Those numbers are response labels, not scores, and they run in the opposite direction to the scoring for positive items. Recode every response before adding anything up.

Which items are reverse-scored?

It depends on the item order, and two orders are in circulation:

The safest approach is to reverse by wording, not by number. The five negatively worded statements are:

  • At times I think I am no good at all.
  • I feel I do not have much to be proud of.
  • I certainly feel useless at times.
  • I wish I could have more respect for myself.
  • All in all, I am inclined to feel that I am a failure.

A worked example

Using the "satisfied with myself" order, a client answers items 1 to 10 as: disagree, strongly agree, agree, agree, disagree, strongly agree, agree, strongly agree, disagree, disagree.

Reversing items 2, 5, 6, 8 and 9, the item scores are 1, 0, 2, 2, 2, 0, 2, 0, 2 and 1, a total of 12 out of 30.

Apply the other key (3, 5, 8, 9, 10) to the same form by mistake and the total becomes 18. A six-point error is larger than the standard deviation in adult norms, and it is in the flattering direction: a mismatched key inflates the scores of clients with low self-esteem.

Is the Rosenberg score out of 30 or 40?

Both conventions are legitimate. The University of Maryland describes 0 to 3 coding (a range of 0 to 30) and notes that you can assign 1 to 4 instead, giving a range of 10 to 40. It adds that some researchers use 5-point or 7-point response scales, which change the range again.

With four response options, the two totals differ by exactly 10. The worked example is 12 on the 0 to 30 range and 22 on the 10 to 40 range. The midpoints are 15 and 25.

This matters whenever you compare a score to published figures. A total of 22 is close to the adult average on one coding and well below it on the other. Always record the range next to the total.

You may also see the RSES called a Guttman scale. That was Rosenberg's original design, and the APA copy describes the method: combined ratings that collapse the 10 items into six scored items. The University of Maryland notes that it is now usually scored as a Likert scale, which is the method described here.

What is a normal Rosenberg Self-Esteem Scale score?

There is no single norm table. The University of Maryland's advice is to look in the research literature for samples similar to your own. Two references are useful starting points:

  • US adults: Sinclair et al. (2010) presented normative data from 503 US adults. The overall mean was 22.62 (SD 5.80) on the 0 to 30 range, and scores differed significantly across age, ethnicity, education, employment, income and marital status groups.
  • Australia and 52 other nations: Schmitt and Allik (2005) gave the RSES to 16,998 people in 53 nations, coded 1 to 4. The overall mean was 30.85 (SD 4.82). The Australian sample of 485 averaged 31.07 (SD 5.15), which is about 21 on the 0 to 30 range. Most samples were university students, so treat it as a reference point, not a population norm.

Every nation in that study averaged above the scale's midpoint. In other words, the typical respondent does not score in the middle of the range. An "average" result sits around 21 to 23 out of 30, not 15.

Back to the worked example: a total of 12 (22 on the 1 to 4 coding) is roughly 1.8 standard deviations below both the US adult mean and the Australian sample mean. That is a more defensible statement in a report than an unreferenced label such as "low".

Schmitt and Allik also found that negatively worded items were read differently from nation to nation, with a neutral response bias more common in collectivist cultures. Be cautious when comparing a client's score to norms from a different cultural or language background.

Is there a cut-off for low self-esteem?

Not an official one. The University of Maryland's Rosenberg Scale FAQ is direct: "There are no discrete cut-off points to delineate high and low self-esteem."

You will still see a rule of thumb that scores below 15 (on the 0 to 30 range) indicate low self-esteem. Wikipedia's entry repeats it, for example. One empirical source for a line in that position is Isomaa et al. (2013), who followed 2,070 Finnish 15-year-olds and suggested a cut-off of 25 points for classifying low self-esteem in both girls and boys. That study used the 10 to 40 range, so 25 there is the same line as 15 on the 0 to 30 range.

Keep the context in mind. The cut-off came from adolescents and was tested against depression and social phobia. The authors presented the RSES as a candidate screening tool that deserved further study, which is not the same as a validated threshold for adult clients.

Low self-esteem is also not a diagnosis. A low RSES score is information for your formulation and a possible treatment target. It does not establish a disorder.

Does the Rosenberg scale have subscales?

It was designed to give one global score. Factor analyses often find that the positive and negative items separate into two factors, and the University of Maryland notes that studies have supported both a one-factor and a two-factor structure.

The strongest test is Marsh, Scalas and Nagengast (2010), who compared competing models using four waves of data across eight years. They found good support for one substantive self-esteem factor plus stable response-style factors linked to item wording. Put simply, a gap between a client's positive-item and negative-item scores may say more about how they respond to negatively phrased questions than about two kinds of self-esteem.

Some research, including Sinclair et al., also reports self-competence and self-liking components. In practice, report the total as the primary score and treat any split as secondary and descriptive.

How should you document and track RSES scores?

An RSES entry that another clinician can interpret records:

  • The item order used and which items were reversed
  • The coding (0 to 3 or 1 to 4) and the resulting range
  • The total and the date
  • The reference sample you compared it with
  • Any unanswered items. The University of Maryland guidance gives no rule for missing responses, so the simplest fix is to ask the client to complete the form

Because the RSES is trait-like, with test-retest correlations in the .80s, expect it to move more slowly than symptom measures. Avoid over-reading a shift of a point or two. It is more informative at review points, alongside a K10, DASS-21 or PHQ-9, than as a weekly tracker.

Self-esteem targets often appear in a depression treatment plan or an eating disorder treatment plan. When the score goes into a report, see our guide to writing a psychological assessment report.

Reverse-scoring by hand is where most RSES errors happen. PractaLuma is AI-native practice management software for Australian mental-health practices. Its standardised assessments let you send the RSES to a client, score it automatically with the correct items reversed, and track results in the client record alongside other measures. See pricing for plan details.

Frequently asked questions

Is the Rosenberg Self-Esteem Scale free to use? Yes. The University of Maryland states that the scale is in the public domain and may be used without charge and without notifying the department.

What is a good score on the Rosenberg Self-Esteem Scale? There is no official "good" score. In the samples above, averages sit around 21 to 23 on the 0 to 30 range (31 to 33 on the 10 to 40 range). Higher scores indicate higher self-esteem.

How do you convert a 10 to 40 score to a 0 to 30 score? Subtract 10. This only works when the scale was given with four response options. Versions with five or seven options have different ranges and are not directly comparable.

What age group is the scale for? It was developed with high school juniors and seniors and has since been used widely with adults. For younger children or specific clinical groups, look for studies using a similar sample.

This guide summarises published scoring conventions for clinicians. It is not a substitute for clinical judgement.