Publication:

Connected speech accuracy in native Nepali-speaking preschool children: a comparative study of typically developing children and children with speech sound disorders

Background Speech sound disorders (SSD) are among the most prevalent communication disorders in preschool-aged children; however, empirical data characterizing connected speech performance in native Nepali-speaking children with SSD remain absent from the international literature. Connected speech analysis offers an ecologically valid complement to conventional single-word assessment paradigms, yet no published normative or comparative framework exists for this population. Objective was to compare connected speech accuracy, operationalized through the percentage of syllables correct (PSC) metric, between typically developing children (TDC) and children with SSD among native Nepali-speaking preschoolers aged 4–5 years.

Methods A comparative cross-sectional design was employed. Participants (n=30) were native Nepali-speaking children aged 4;0 to 5;00 years recruited from a preschool (TDC; n=15) and clinical setting (SSD; n=15) in Kathmandu. Connected speech samples were elicited via a structured sentence repetition task based on the culturally familiar "thirsty crow" narrative, supported by ten illustrated picture stimuli. PSC was calculated from broad phonetic transcription conducted by a qualified speech-language pathologist with expertise in Nepali phonology. Inter- and intra-rater reliability were established. Group differences in PSC were examined using the Mann–Whitney U test following confirmation of non-normality via the Shapiro–Wilk test.

Results TDC achieved a mean PSC of 93.58% (SD=6.65), reflecting near-ceiling syllable accuracy. Children with SSD demonstrated substantially reduced performance (mean PSC=63.14%; SD=8.62). The Mann–Whitney U test revealed a highly significant between-group difference (U=0.000, Z=−10.887, p<0.001), with complete absence of distributional overlap.

Conclusions PSC derived from a structured connected speech task demonstrated high discriminative sensitivity in distinguishing TDC from children with SSD in a Nepali-speaking preschool population. These findings support routine integration of connected speech analysis into early speech-language assessment and provide preliminary normative data for an underrepresented linguistic group.