Reliability and Concurrent Validity of the Likert-Type Stuttering Severity Scale and the Visual Analog Scale in Turkish-Speaking Adults Who Stutter.

Fifty-two final-year speech-therapy students in Ankara (26 per group) listened to video-recorded speech samples from 15 adults who stutter (aged 18–32) and rated severity using either an 11-point numerical scale (0 to 10) or a 100-mm visual analog line. Raters were randomly assigned to a scale, stratified by gender. Each group received a brief orientation with three anchor recordings and two practice samples. Ratings were completed online across three sessions in one day; a one-week interval separated the test and retest, during which five samples were re-presented. Two experienced clinicians independently scored all recordings with the SSI-4-TR to serve as the reference. Raters were blinded to the clinician scores. Significance was set at p < 0.01.

Test-retest reliability was excellent for both tools (VAS ICC .997; SEV ICC .987). Inter-rater reliability was moderate, with the VAS slightly ahead (ICC .656 vs .573). Both scales correlated strongly with clinician SSI-4-TR scores (VAS r = .835; SEV r = .814), indicating that trained non-clinician perceptual ratings closely tracked the standard clinical measure. No missing data; all 52 raters completed all sessions.

If you need a quick severity estimate for screening, a large research cohort, or a setting where a full SSI-4 is impractical, a trained non-clinician (e.g., a senior student or research assistant) can rate stuttering severity on a simple 0-to-10 scale or a 100-mm line and get a result that tracks closely with your own SSI-4 score. However, two different raters may still disagree meaningfully (inter-rater ICC only .573–.656), so do not rely on a single non-clinician rating for treatment decisions. The tools are best used as a supplement to, not a replacement for, the SSI-4.

All 52 raters were final-year speech-therapy students with supervised clinical exposure to fluency disorders; the findings may not extend to truly untrained listeners or to raters from other disciplines. Only 15 speakers were used, all adults (18–32), all native Turkish speakers; the results say nothing about children, older adults, or speakers of other languages. All ratings were done online via personal computers and headphones, so audio quality and listening conditions varied across raters. The study measured correlation with SSI-4-TR scores, not agreement; a strong correlation does not mean the two tools produce interchangeable numbers for a given patient. Inter-rater reliability was only moderate, meaning two raters using the same tool may assign meaningfully different severity levels to the same speech sample.

The authors declare no conflicts of interest and state the research did not receive support from any funding agency or industry.

Do not read the strong correlations (r = .835 and .814) as permission to drop the SSI-4 in favour of these quick scales for clinical decision-making. All 52 raters were trained speech-therapy students with supervised clinical exposure, not untrained listeners, and inter-rater agreement was only moderate (ICC .656 and .573), so two raters may meaningfully disagree on the same speech sample.