General Self-Efficacy Scale: 10 statements, scored 10 to 40
Self-efficacy is the belief that your own actions will make a difference to what happens next. Schwarzer and Jerusalem built the General Self-Efficacy Scale in 1995 to measure it in ten statements, and the German original dates to 1979.
Each statement is rated 1 (not at all true) to 4 (exactly true). Nothing is reverse-scored, so the total runs from 10 to 40 and takes about four minutes. The scale is in the public domain: the authors gave up copyright, so it can be used and translated freely.
One factor across many countries
Ten items, one factor. Scholz and colleagues tested it in 25 countries with about 20,000 participants and the single-factor structure held in most country samples. A Ugandan sample of community health workers extracted three factors (problem solving, persistence, resourcefulness) with an alpha of .61, so the one-factor answer is not universal.
| Country | Mean |
|---|---|
| Japan | 20.22 |
| Hong Kong | 23.05 |
| United States | 29.5 |
| Germany | 29.3 |
| France | 32.19 |
| Denmark | 32.87 |
| Costa Rica | 33.19 |
The international mean was 29.55 (SD 5.2). Japan sits lowest and Costa Rica highest, and a spread of 13 points across countries is wide enough that a raw score means little without knowing who you are being compared against. Chinese samples tend to land slightly below Western ones, in the 26 to 30 range for students.
Clinical groups sit lower as well: 32.11 for people with coronary heart disease, 30.63 for cancer patients, 28.29 for advanced cancer, 30.0 for people living with HIV, 28.77 for adults without an injury, 29.05 and 28.71 for older and younger arthritis patients, and 22.4 and 20.8 for people with schizophrenia in supported housing.
Reliability
Internal consistency across the 23-country sample ranged from .76 to .90, with most countries above .85. Split-half reliability was .80 to .85. Test-retest over six months ranged from .55 to .75, and over a year from .47 to .63. Those are moderate figures, and they fit the construct: efficacy beliefs move when circumstances move, and a scale that ignored that would be measuring something else.
Chinese versions report .85 to .91 in students, .83 to .88 in community adults and .80 to .86 in clinical samples, with a four-week retest of .72 in the student data.
What it correlates with
| Measure | r |
|---|---|
| Optimism | .50 to .60 |
| Self-esteem | .50 to .60 |
| Social support | .30 |
| Depression | -.40 to -.55 |
| Anxiety | -.35 to -.50 |
| Perceived stress | -.30 to -.45 |
| Burnout | -.35 to -.50 |
Self-esteem overlaps enough to matter: the two share roughly 25% to 36% of their variance, so a high score here is partly a high score there. The negative correlations with depression and anxiety are moderate, which is what you would expect from a belief about coping rather than a measure of mood.
A six-item short form was built from 19,719 people across 26 countries by keeping the most discriminating items. It reports alphas of .79 to .88, correlates -.35 to -.45 with depression, -.35 with anxiety and -.38 with exhaustion, and predicts both mental and physical health 28 months later after controlling for baseline.
Limits
Ceiling effects show up in Western samples, where many people score above 35, so the scale separates high scorers poorly. General self-efficacy is stable but blunt: a domain-specific measure of academic or social efficacy will usually predict domain-specific behaviour better. And the items lean on personal agency, which may not translate cleanly into collectivist settings.
There is no cutoff. The authors suggest using a sample median, or one standard deviation either side of it, for grouping. That is a research convention rather than a clinical line. A total is not a diagnosis, and this page does not replace one.