Eleven /s/-lateralizing kindergarten children were followed for two years during kindergarten and first grade, and their fricatives analyzed perceptually. Their kindergarten misarticulations occurred only on /s/ or, at most, one other sound. The present authors were intrigued when it was discovered that eight of 11 original lateralizing kindergarteners, who were now in 10th grade in a local high school, were still lateralizing /s/. Five of the eight consented to be reassessed. Each had received 4-5 years or more of conventional speech therapy after they had not spontaneously overcome lisping in first grade. For four of the five subjects, acoustic measures of lateralization had decreased, and for one, increased (Stephens et al., 1999a; Stephens, Lu, Kao, Khavazadeh & Daniloff, 1999b). Spectral moments for the five subjects revealed context sensitivity that was highly idiosyncratic across speakers and contexts and that had changed over time. We speculated that different spectral measures might yield fresh insight into context sensitivity, thus yielding information that might be used profitably for conventional articulatory treatment. These children represented the “discrete” misarticulators noted by Gibbon, Hardcastle, and Dent (1995), and would therefore be likely to profit from EPG training.