Anxiety is not one thing
Two people with the same diagnosis can describe almost nothing in common. One cannot stop worrying about money. The other feels clear-headed but has a racing heart, shaking hands and a stomach that knots before every meeting. Anxiety runs as thoughts for some people and as body symptoms for others. The Zung Self-Rating Anxiety Scale (SAS) was built to catch both, which is also why it does not sort anxiety into types. What it gives you is a severity reading: how much anxiety is present right now, taken as a whole.
What the 20 items ask
William Zung published the scale in 1971 in Psychosomatics. Each of the 20 statements describes something from the past week, and you rate how often it applied: none or a little of the time, some of the time, a good part of the time, or most or all of the time.
Five items are worded in the calm direction ("I feel calm and can sit still easily", "My hands are usually warm and dry", "I fall asleep easily and get a good night's rest"). They are reversed before scoring, so answering them in the calm direction does not push the total up. The raw total runs from 20 to 80, and multiplying it by 1.25 gives the index score that most clinical papers report.
Ten of the twenty items are about the body
Trembling arms and legs, headaches and neck pain, weakness and fatigue, a pounding heartbeat, dizzy spells, fainting spells, numbness in the fingers and toes, stomach trouble, frequent urination, a flushed face. Four items cover the psychological side (feeling nervous, afraid without reason, panicky, like falling apart), two ask about sleep, and four are the reversed calm items.
Zung wrote the scale for medical and psychiatric outpatients, and that is still where it gets used most. Covering the body has a cost: a physical condition moves the score. Someone with an untreated thyroid problem, chronic pain, or a medication that causes tremor can answer the somatic items honestly and land higher than their anxiety alone would justify.
Three different cutoffs, and why
Zung's original threshold was a raw score of 40, which is an index score of 50. Later work argued that this misses too many people when the scale is used for screening. Dunstan and Scott (2020, BMC Psychiatry) checked both thresholds against PHQ-based diagnoses in a clinical sample of 141 people and a community sample of 210.
At a raw score of 36, the scale caught 89% of the cases and flagged 25% of people who did not meet the reference diagnosis. At the original 40 it caught 75% and flagged 10%. Their recommendation: use 36 for clinical screening, where a missed case costs more than an extra follow-up, and 40 for research, where a cleaner sample matters more. A Chinese study of 1,109 university students used an index score of 45, which comes to the same raw score of 36.
This page reports four bands, following the scoring sheet published by Florida State University's Institute for Trauma and Resilience Studies:
| Raw score | Reading |
|---|---|
| 20 to 44 | Normal range |
| 45 to 59 | Mild to moderate anxiety |
| 60 to 74 | Marked to severe anxiety |
| 75 to 80 | Extreme anxiety |
These bands sit on the raw total, not the index score, and they are a reading aid rather than a clinical line. For scale: a clinical sample of people who did not meet the PHQ threshold averaged 36.21 (SD 6.50), and the Chinese student sample averaged 44.56 (SD 8.64).
How well it works
Reliability is decent without being exceptional. Cronbach's alpha was 0.78 in the Chinese student sample and 0.80 in a study of adults with intellectual disability (Ramirez and Lukenbill, 2008). The SAS correlates with the Beck Anxiety Inventory at 0.54 to 0.66: high enough that the two measure related things, low enough that they are not interchangeable.
The same Chinese study is the source of the honest weakness. In an item response theory comparison, the SAS gave less information than the Beck Anxiety Inventory and the State-Trait Anxiety Inventory at every level of severity. The five reversed items were the weakest part of the scale, with poor category functioning. In a Vietnamese inpatient sample with recurrent depression, the SAS also struggled to separate severe depression (area under the curve 0.5923), though it was the most stable of the four scales tested in where the best cutoff fell, staying between 44 and 46 across subgroups.
Take the test
The CheckPsych version uses the 20 original items and takes about 7 minutes. You get the raw total, the index score and one of the four bands above.
If the somatic items are the ones you endorsed, get a medical check before reading much into the number. Several physical conditions produce the same pattern, and a clinician can tell them apart.
References
Zung, W. W. K. (1971). A rating instrument for anxiety disorders. Psychosomatics, 12(6), 371-379.
Dunstan, D. A., & Scott, N. (2020). Norms for Zung's Self-rating Anxiety Scale. BMC Psychiatry, 20, 90.
Pang, Z., Tu, D., & Cai, Y. (2019). Psychometric properties of the SAS, BAI, and S-AI in Chinese university students. Frontiers in Psychology, 10, 93.
Ramirez, S. Z., & Lukenbill, J. (2008). Psychometric properties of the Zung Self-Rating Anxiety Scale for adults with intellectual disabilities. Journal of Developmental and Physical Disabilities, 20(6), 573-580.
Tung, V. S., et al. (2023). Diagnostic value in screening severe depression of the HAM-D, HAM-A, BDI and Zung's SAS among patients with recurrent depression disorder. Acta Informatica Medica, 31(4), 249-253.
Disclaimer: this article and the SAS are for education and self-reflection. They are not a diagnosis. Anxiety scores move with sleep, caffeine, illness and whatever else is going on that week. If your score lands in the marked or extreme range, or if anxiety is interfering with work, sleep or relationships, talk to a qualified professional.