Purpose: Prior studies have shown that bilingual children who do not stutter (CWNS) exhibit a high number of disfluencies in both languages, increasing the risk of misidentification by speech-language pathologists as children who stutter (CWS). Conversely, there is a risk of misidentifying CWS with a relatively low incidence of disfluencies as CWNS. This study aims to explore the qualitative and quantitative distinctions in speech disfluency profiles between CWNS and CWS. The assessment covers both the dominant and nondominant language to examine the impact of language dominance on disfluency patterns.

Method: A total of 92 Lebanese bilinguals (70 CWNS and 22 CWS) from 4;06 to 7;06 (years;months) were included. Language dominance was determined based on parental assessments. Spontaneous and narrative speech samples were collected for each child in both languages and all stuttering-like disfluencies (SLD) and other disfluencies (OD) were coded.

Results: On average, CWNS showed a significantly lower percentage of total SLD, weighted SLD, SLD subtypes, and iterations compared to CWS. However, the number of disfluencies of CWNS exceeded monolingual clinical standards. Language dominance did not impact SLD and OD percentages, but some differences for SLD subtypes emerged. Binary logistic regression analyses showed that repetitions and dysrhythmic phonations are good predictors for correct CWS or CWNS classification, in contrast to OD. A combination of predictors from both languages led to better classification than using predictors from either language alone.

Conclusions: The current study shows that speech disfluency percentages in bilingual CWNS typically surpass monolingual standards and can be at par with those of CWS. However, through careful consideration of disfluency characteristics, ideally in both languages, an accurate differential diagnosis of stuttering in bilingual children can be achieved.