Anxiety Tests Explained: BAI vs. GAD-7
When, why, and how to use two of the most common anxiety scales.
The Beck Anxiety Inventory (BAI) and the Generalized Anxiety Disorder-7 (GAD-7) are the two instruments most often reached for when anxiety is the presenting concern. Both are described as anxiety measures, and that shared label hides a real difference: they answer different questions. A well-administered but poorly chosen instrument can manufacture a problem that is not there, or miss one that is.
Two scales, two different questions
The BAI asks: how intensely is anxiety being felt in the body right now? The GAD-7 asks: does this person's worry pattern look like generalized anxiety? One measures severity; the other screens for a syndrome. Every difference below follows from that distinction.
Structural comparison
| Property | BAI | GAD-7 |
|---|---|---|
| Number of items | 21 | 7 |
| Score range | 0–63 | 0–21 |
| Centre of gravity | Somatic symptoms | Worry and cognitive symptoms |
| Typical administration time | 5–10 minutes | 2–3 minutes |
| Primary use | Grading symptom severity | Rapid screening and repeated monitoring |
| Weakest at separating | Anxiety from medical causes of the same bodily signs | Generalized worry from panic or social anxiety |
When the BAI earns its length
If the bodily component dominates the presentation, the BAI produces detail the GAD-7 cannot. Seeing palpitations, sweating, trembling, and breathlessness rated separately helps in two ways: it sharpens the treatment target, and it gives the person a vocabulary for something they have been experiencing without a name. For someone who opens with "I am not anxious, my heart just races," the BAI is often the fastest route to a shared language.
When the GAD-7 is the better tool
Brevity here is more than convenience. A seven-item instrument can be completed without derailing a first session, sent out every fortnight, and repeated without fatiguing the person answering it. If the goal is tracking response to treatment, screening in a primary-care or workplace setting, or collecting pre- and post-measures in a group programme, the GAD-7 is the pragmatic choice.
The somatic trap
Because BAI items concentrate on bodily sensations, they overlap with a long list of non-psychiatric causes: thyroid dysfunction, anaemia, arrhythmias, heavy caffeine intake, and the side effects of several common medications all produce the same sensations. Someone in that position can post a high somatic anxiety score without having an anxiety disorder at all. A high BAI does not substitute for medical evaluation; it frequently calls for one.
The reverse trap is quieter. A years-long, life-narrowing worry pattern that generates little in the way of bodily symptoms can land in the low range. Reading that as "low score, no problem" is a mistake.
A case: high on one, low on the other
A 34-year-old presents with "I am tense all the time." The GAD-7 comes back high: uncontrollable worry, restlessness, and difficulty concentrating are all endorsed strongly. The BAI sits in the mild range.
That combination is informative on its own. When the worry load is heavy but the somatic load is light, the early targets are usually cognitive — working with worry itself, tolerance of uncertainty, and postponed decisions. Reversed, with a high BAI and a low GAD-7, priority would move toward body-based regulation and a medical differential. The difference between the two scores carries more information than either alone.
Which tool for which situation
| Situation | First choice | Why |
|---|---|---|
| Rapid screen at intake | GAD-7 | Short enough not to disrupt the session |
| Grading bodily symptom severity | BAI | Somatic items are itemized in detail |
| Fortnightly progress monitoring | GAD-7 | Repeats easily without burden |
| Suspected panic | BAI plus a panic assessment | Episode structure has to be asked about separately |
| Anxiety confined to social settings | Social anxiety scale | General anxiety measures do not distinguish context |
| Anxiety with low mood | GAD-7 plus a depression measure | Overlap is common; one instrument narrows the picture |
The reporting window: the detail most often skipped
Administered in the same session, the two instruments should not be expected to agree exactly, because they do not ask about the same stretch of time. The BAI is anchored to recent days; the GAD-7 covers a two-week window. Someone who had an intense episode last week but a calm few days since will produce a predictable divergence. For comparisons across time to mean anything, both the instrument and the window have to be held constant.
Anxiety rarely arrives alone
When an anxiety measure comes back elevated, the next question is always what else is present. Overlap with depression is so common that screening for only one is a routine omission. Where obsessional features stand out, the Padua Inventory adds discriminating information; where the tension is tied to sustained workload, a burnout measure separates two pictures that look alike but call for different responses. If concentration problems are part of the complaint, the ASRS helps establish whether they belong to the anxiety or stand on their own — both produce the same symptom, and the interventions differ.
The shared limit: neither one diagnoses
Both instruments screen and grade. An elevated score establishes that symptoms are being reported intensely; it does not establish an anxiety disorder. Duration, functional impact, and the exclusion of other explanations remain the work of clinical assessment.
Frequently asked questions
Is administering both redundant?
No. They measure different components, so together they are complementary. If time is short, start with the GAD-7 and add the BAI when the interview suggests a meaningful bodily load.
The two scores disagree. Which one do I trust?
The disagreement is itself the finding. Rather than picking a winner, ask what the gap describes: is the load somatic or cognitive?
Can these be used with adolescents?
Both were developed for adults. For children and adolescents, age-appropriate instruments such as the SCARED or the RCADS-25 are preferred.
How often should they be repeated?
Intervals of two to four weeks are common in monitoring. More frequent administration mostly captures measurement fluctuation rather than real change; much longer gaps delay noticing whether treatment is working.
For the rules that govern reading either score, see what a psychological test score actually means. For sending these instruments out between sessions, see the digital test administration guide.
This article is for information only. The scales and tests mentioned here are screening and assessment instruments; on their own they do not establish a clinical diagnosis. Administration and interpretation belong to licensed professionals trained in the relevant instrument. Item texts, stimulus cards, and norm tables of copyrighted tests are never published on this page.