Is it necessary to assess fluent symptoms, duration of dysfluent events, and physical concomitants when identifying...

Riley's (1994) Stuttering Severity Instrument (SSI) has three components: a symptom frequency measure (%SS), average duration of the three longest stutters, and a physical concomitant (PC) score. An assessment of whether it was necessary to use all of these when using SSI-3 to identify which children are at risk of speech difficulties was performed. Participants were 879 reception class children aged 4-6 years from UK schools. The distributions of the separate components of SSI-3 were examined. Departures from normality were noted for each component. The features seen in the distribution of the individual components were also apparent in the distribution of the overall scores (this was not normal and had multiple modes). These findings undermine the usefulness of the overall measure for identifying children at risk of speech difficulties. Prior work used a fixed SSI-3 threshold to identify at risk children. Classification of children as fluent or at risk based on this threshold was compared with classifications based on thresholds applied to the individual components. Classifications were comparable for %SS, but less satisfactory for duration and PC. These findings suggest that %SS performs similarly to the overall SSI-3 scores when used to identify at risk children. Procedures for identifying at risk children in schools need to be short and easy to administer. Thus, since there is no justification for including all components of SSI-3 and duration and physical concomitants are not sensitive measures of fluency, a procedure based on the frequency measure alone is appropriate for use in schools.