Social Phobia Inventory (SPIN):
Scoring and Interpretation

The SPIN is a 17-item self-report measure developed to capture social anxiety symptoms, including fear, avoidance and physiological manifestations. Connor et al. (2000) sought a brief instrument that covered all three types of experience relevant to social phobia, now generally described as social anxiety disorder.

The SPIN has been studied as both a screening measure and a symptom-change endpoint in social anxiety research. Its practical advantage is a simple 0 to 68 total. However, a total above a research threshold is a positive screening result, not an individual clinical diagnosis.

• Items: seventeen symptom ratings, each scored 0 to 4, with no standard reverse-coded items.
• Recall period: the past week in the standard self-report form.
• Total: the 17 responses are summed to a 0 to 68 total; higher totals mean greater social anxiety symptom burden.
• Subscales: the original developers investigated symptom facets, and the original factor analysis identified five factors, but no universal five-subscale scoring procedure is implied for routine clinical use. The commonly reported primary score is the 17-item total.
• Short form: do not conflate the full 17-item SPIN with the three-item Mini-SPIN, a separate short screener.
• Translations: translated and age-group adaptations require separate verification.

1. Verify that the administered instrument contains all 17 SPIN items coded 0 to 4 and in the original order (spin1 to spin17).
2. Sum all 17 responses to form spin_total, which ranges from 0 to 68. No reversal is needed.
3. Optionally define spin_pos for a total of 19 or more, if using the original overall clinical-versus-control screening comparison.
4. Use complete-case scoring in the sample. The cited sources set no single prorating rule for incomplete forms, so pre-specify any other procedure.

* Illustrative syntax: variables must already be valid item scores (0 to 4).
COMPUTE spin_total=SUM.17(spin1,spin2,spin3,spin4,spin5,spin6,spin7,spin8,spin9,spin10,spin11,spin12,spin13,spin14,spin15,spin16,spin17).
EXECUTE.
IF (NOT MISSING(spin_total)) spin_pos=(spin_total>=19).
EXECUTE.

# data: one row per respondent; items spin1 to spin17 coded 0 to 4.
data$spin_total <- rowSums(data[paste0("spin",1:17)],na.rm=FALSE)
data$spin_pos <- ifelse(is.na(data$spin_total),NA,data$spin_total>=19)

In the original clinical social-phobia versus pooled-control comparison, a score of 19 produced sensitivity 0.725 and specificity 0.843. Different comparison samples yielded different optimal thresholds (15 against nonpsychiatric controls and 16 against psychiatric non-social-phobia controls), illustrating why 19 is not universal.

For treatment studies, the 0 to 68 score can be a continuous primary symptom outcome. Use a baseline-adjusted Multiple Linear Regression, or Linear Mixed Models (Multilevel Modeling) across repeated visits, standardize the recall period and consider baseline severity when interpreting group differences.

For screening datasets, describe the threshold and its source comparison group, then report Sensitivity & Specificity (Diagnostic Accuracy) with confidence intervals only if you have a diagnosis or structured interview reference, and use ROC Curve Analysis to examine alternative thresholds in that sample. Reliability estimated in the current sample is preferable to quoting original studies alone, and the factor structure of a translated version can be examined with Exploratory Factor Analysis (EFA). No universal minimal clinically important difference is established in the original abstract.