In reply
Drs St John and Nelson discuss a methodological concern of considerable importance to cochlear implant research: quantifying levels of speech recognition in development. They correctly quote the prior report by Cheng et al1 noting that the speech perception category score has not been shown to be parametric and so these scores must be considered ordinal data for statistical analysis. Our use of t tests to analyze these data may therefore be considered problematic. Nonetheless, this criticism does not alter the validity of the principal results of the article, including the demonstrated improvement in speech perception scores subsequent to implantation. This can be seen from the original data in the article and is confirmed by a nonparametric analysis that was unfortunately omitted from the article.