Home / Resources / Psychological Measures / BAI
Beck Anxiety Inventory:
Scoring and Interpretation
The Beck Anxiety Inventory (BAI) is a 21-item self-report scale of anxiety symptom severity over the past week, weighted toward physical and subjective symptoms and scored from 0 to 63 (Beck et al., 1988). Coverage begins with Beck Anxiety Inventory scoring and the reason the 1993 bands must not be mixed with the older 1990 ones, then moves to SPSS and R code and methods wording for the 0 to 63 total.
What the Beck Anxiety Inventory Measures
The BAI was introduced to quantify clinical anxiety, in particular the physical and subjective symptoms that accompany it. Beck et al. (1988) examined its psychometric properties in clinical samples, including how well it separated anxiety symptoms from depressive symptoms.
It suits psychiatric and behavioral-health research in which the analytic target is symptom intensity. Because many items concern bodily sensations, comorbid physical illness has to be weighed when scores are interpreted. The BAI is not a structured psychiatric diagnostic assessment.
Structure of the BAI
- Items: 21 symptom ratings, each coded from 0 to 3 on four severity categories.
- Recall period: the past week, including the day of assessment.
- Subscales: none in standard scoring. A single summed score is the usual reporting convention.
- Total: one score from 0 to 63.
- Reverse-scored items: none.
- Unscored content: there is no standard additional functional-impact item outside the total.
- Translations: validated translated versions exist, and equivalence should be evaluated in the population studied.
The instrument itself is available from Pearson Assessments.
How to Score the BAI
- Check that bai1 to bai21 contain item scores in the range 0 to 3 and not arbitrary response-category codes.
- Add all 21 items to form bai_total, which ranges from 0 to 63. No reversal is applied.
- Decide how incomplete forms are handled. This guide does not apply a publisher-specific rule for incomplete forms, so the sample code requires complete valid responses. If you use another rule, state it in the methods section.
* Illustrative syntax: variables must already be valid item scores (0 to 3).
COMPUTE bai_total = SUM.21(bai1, bai2, bai3, bai4, bai5, bai6, bai7, bai8, bai9, bai10, bai11, bai12, bai13, bai14, bai15, bai16, bai17, bai18, bai19, bai20, bai21).
EXECUTE.# data: one row per respondent; items bai1 to bai21 coded 0 to 3.
data$bai_total <- rowSums(data[c("bai1", "bai2", "bai3", "bai4", "bai5", "bai6", "bai7", "bai8", "bai9", "bai10", "bai11", "bai12", "bai13", "bai14", "bai15", "bai16", "bai17", "bai18", "bai19", "bai20", "bai21")], na.rm = FALSE)Interpreting BAI Scores
| Score band | Interpretation |
|---|---|
| 0 to 7 | Minimal |
| 8 to 15 | Mild |
| 16 to 25 | Moderate |
| 26 to 63 | Severe |
Source: Beck and Steer (1993) BAI manual, as summarized by the Shirley Ryan AbilityLab RehabMeasures Database (n.d.).
Higher totals correspond to more severe self-reported anxiety symptoms, and the four categories are descriptive severity ranges. No single threshold reliably diagnoses a specific anxiety disorder in every patient group, and reported sensitivity and specificity depend on the clinical setting and the reference condition. For BAI interpretation, that means the bands describe symptom load and do not classify cases.
Overlap between somatic anxiety symptoms and physical health problems can raise a BAI total without implying an anxiety disorder. An earlier manual (1990) used different bands (0 to 9 normal, 10 to 18 mild to moderate, 19 to 29 moderate to severe, 30 to 63 severe), and some older articles still apply them, so name the exact scheme you used and never mix the two.
Psychometric Properties
| Property | Published finding | Source |
|---|---|---|
| Internal consistency | Cronbach’s α approximately .92 in the original psychometric research | Beck et al. (1988), summarized in subsequent assessment literature |
| Test-retest | One-week correlation of approximately .75 reported for the original study | Beck et al. (1988), secondary summaries |
| Convergent validity | Correlations with clinician anxiety ratings support convergence; the numerical estimate varies by comparator | Beck et al. (1988) |
| Diagnostic accuracy | No stable universal screening sensitivity or specificity for a single threshold across conditions | Shirley Ryan AbilityLab (n.d.) |
| Subscales | A single summed score is the usual reporting convention | Beck et al. (1988) |
These are source-specific results and cannot be assumed to hold as reliability or accuracy estimates in a new sample.
Using the BAI in a Study
The 0 to 63 total can be modeled as a continuous variable in comparative and longitudinal analyses, with attention to positive skew and boundary effects. With repeated waves, Linear Mixed Models (Multilevel Modeling) keep the within-person structure intact. Decide in advance whether the main scientific target is anxiety symptom burden or case finding, because the two call for different claims and possibly different measures.
Report the internal consistency of your own sample, using Reliability Analysis (Cronbach’s Alpha), together with the distribution of totals. In medically ill populations, assess whether physical symptom burden is a likely confounder and, where the design supports it, adjust for it in a Multiple Linear Regression. If translated versions are compared, test Measurement Invariance before reading differences in means. Power calculations should target the anticipated effect size and model, since no instrument-specific minimum sample applies.
Related field: Psychology & Behavioral Sciences
Reporting the BAI
Anxiety symptoms were assessed with the 21-item Beck Anxiety Inventory (BAI; Beck et al., 1988). Each symptom was rated from 0 to 3 for severity and the ratings were summed (possible range = 0 to 63), with higher totals indicating greater anxiety symptom burden. Internal consistency in this sample was [α = XX].
References
Beck, A. T., & Steer, R. A. (1993). Beck Anxiety Inventory manual. Psychological Corporation.
Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56(6), 893-897. https://doi.org/10.1037/0022-006X.56.6.893
Shirley Ryan AbilityLab. (n.d.). Beck Anxiety Inventory. https://www.sralab.org/rehabilitation-measures/beck-anxiety-inventory
Frequently Asked Questions
Does the BAI diagnose generalized anxiety disorder?
No. The BAI quantifies the severity of anxiety symptoms and cannot take the place of the clinical diagnostic process.
What does a BAI score of 16 mean?
Under the 1993 severity ranges used in this guide, 16 is the lower boundary of the moderate category. Under the older 1990 bands the same total would fall in a different category, which is why the scheme must be named.
Do physical health symptoms matter when interpreting the BAI?
Yes. Somatic overlap can lift the score, so consider it at the design stage and again when you interpret the results, especially in medically ill samples.
Can I compare BAI totals with GAD-7 totals?
Do not equate the raw totals. The two scales assess overlapping but not identical constructs and have different ranges, so any comparison needs an appropriate analytic strategy rather than a direct subtraction.
NEXT STEP
Need BAI Scores Computed
or Reviewed?
OnlineSPSS is a statistical consultancy for researchers. Share your design and dataset, and a statistically trained Research Coordinator reviews it. The scope and one fixed project fee are confirmed before work begins.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014