For longitudinal research how can correlations be computed between the preschool scores and subsequent scores on the school-age forms?

Answer:  Correlational and regression statistics are not affected by the absolute magnitudes of the scores at each assessment point.  Consequently, they can be used to compute longitudinal associations between the raw scores or T scores obtained from each instrument.  For syndrome scales and DSM-oriented scales, raw scores provide more differentiated analyses than T scores do, because the T scores are truncated at 50.  For the Internalizing, Externalizing, and Total Problems scales, T scores can be used with no loss of differentiation, because they are truncated at 50. To provide a more precise measure of longitudinal associations between the corresponding sets of items on the preschool and school-age forms, correlations can be computed between the sum of scores on the CBCL/1.5-5 items that have counterparts on the CBCL/6-18 and the sum of scores on their counterparts for older ages.  Likewise, correlations can be computed between the sum of scores for the items that have counterparts on the C-TRF and TRF.