Individual differences in language and working memory affect children's speech recognition in noise.

The children were all typically developing, had normal hearing and were native English speakers from English-speaking homes, so the findings may not apply to children with hearing loss, other disabilities or different language backgrounds. Some children could not complete the sentence and sequence tasks: 26 were not analysed for low-predictability sentences and 33 were not analysed for zero-predictability sequences, which may make those conditions harder to use clinically. Working memory measures were highly correlated, so the study could not tell whether simple span, complex span, verbal or visuospatial working memory mattered more, or whether a general thinking ability was driving the result. Vocabulary, working memory and syntax shared variance, and no intelligence measures were included, so the unique contribution of each predictor was hard to separate. The study used steady-state masking noise based on one talker, not competing talkers or reverberation, and tested children in one session, so it shows associations rather than everyday classroom listening or causation.

The authors reported no conflicts of interest. The study was supported by grants from NIH-NIDCD.

Do not conclude that poor speech recognition in noise is caused by weak working memory or limited vocabulary. This was a single-session measurement study in children with normal hearing, and it reported associations rather than treatment effects.