Depression Anxiety Stress Scales-21:
Scoring and Interpretation

The DASS-21 assesses three related dimensions of recent emotional distress: depressive mood, anxiety-related arousal and stress-related tension. It is the 21-item short form of the original 42-item DASS and produces three seven-item scores (Lovibond & Lovibond, 1995).

In a large nonclinical adult sample (N = 1,794), Henry and Crawford (2005) examined its measurement structure and found support for specific depression, anxiety and stress dimensions alongside a general distress factor. The DASS-21 therefore suits research on several kinds of distress at once, although its subscales are not stand-alone psychiatric diagnoses.

• Depression items: 3, 5, 10, 13, 16, 17 and 21.
• Anxiety items: 2, 4, 7, 9, 15, 19 and 20.
• Stress items: 1, 6, 8, 11, 12, 14 and 18.
• Response scale: four options coded 0 to 3, with a one-week recall period.
• Scores: each raw subscale total runs from 0 to 21, and the doubled DASS-42-equivalent version runs from 0 to 42. There are no reverse-scored items and no standard fourth subscale.
• Age range: adults and adolescents from 14 years. The DASS-Y is a distinct youth version recommended for younger respondents (University of New South Wales, n.d.).

1. Validate dass1 to dass21 as integers from 0 to 3.
2. Sum the seven items of each subscale to obtain dass_d_raw, dass_a_raw and dass_s_raw. Each raw total ranges from 0 to 21.
3. Multiply each raw total by 2 to obtain the DASS-42-equivalent scores (dass_d_42, dass_a_42, dass_s_42) before applying the original severity tables. Never double the scores and then treat them as if they ran from 0 to 21.
4. Choose a missing-item rule and state it. The developer FAQ (item 27) says one missing item can be handled by averaging the remaining items in that scale, with a rule of thumb of at most one missing item per seven-item DASS-21 scale (two per 14-item DASS-42 scale); beyond that, treat the scale as invalid. The code below requires all seven items, so add prorating if your protocol allows one missing item.

Larger values indicate more frequent or intense symptoms in the corresponding domain. The severity labels are population-relative: the developers describe the cut-points as arbitrary divisions of the full score range, so “mild” means above the population mean but well below the typical severity of people presenting for treatment, not a mild disorder. They are not diagnoses and do not establish a psychiatric disorder. When you report DASS-21 interpretation, state whether the values are raw seven-item totals or doubled DASS-42-equivalent scores.

Analyze the Depression, Anxiety and Stress totals separately when they correspond to distinct hypotheses. For group or time comparisons the raw 0 to 21 scores are convenient. Doubling is a linear rescaling, so correlations, standardized effects and test statistics for a linear contrast stay the same, while unstandardized differences and their confidence limits double. Apply the published bands only to correctly doubled scores.

Report reliability for each subscale in your sample, as alpha or omega, rather than a single alpha across all 21 items; Reliability Analysis (Cronbach’s Alpha) covers the calculation. The three scores share a general distress component, so inspect their correlations before entering all three into one Multiple Linear Regression. In factor-analytic work, such as a Confirmatory Factor Analysis (CFA) of the general-plus-specific structure, base the sample size on the parameterization, the precision needed and the study objectives. Comparisons across language versions or age groups call for Measurement Invariance testing before means are interpreted.