In Germany, international and national comparative studies such as PISA or IGLU have shown that around 20% of school children do not reach the minimal competence level as defined by PISA literacy stages by the age of 15. These children usually do not reach University. One indicator is the ability to spell correctly but the effort involved in detailed analysis in this area has rendered longitudinal studies laborious to impossible. With the advent of automated detailed error profiling this may change. Lacking extensive data, the work presented here focuses on a long-term case study evaluated against the background of available cross sectional data from a previous study in order to understand the mechanics needed for automated individual diagnostics and use know-how in the literature to track a specific child's progress. We are able to show that a meaningful data representation can be found for quick diagnostic feedback. However, extended collection and evaluation of longitudinal data is necessary even beyond 6th grade to enable comprehensive feedback.

Index Terms: speech synthesis, spelling errors, case study, education, German