Thanks for asking. Sorry I don't have a visual display prepared, but I will attempt to provide a descriptive example here which I hope will clarify things a bit. First of all, yes, I would definitely suggest using a regression model, but the actual model used doesn't really matter. I am merely suggesting that you start with whatever discrepancy model you currently use to find the severe discrepancy value then apply a band of plus or minus 5 points to that value. For example using the Minnesota regression formula, assuming that the subject earned an IQ score of 100, the severe discrepancy value for an achievement score would be 79. Applying a band of plus or minus 5 points to that value any achievement score falling between 74 and 84 would fall within the SLD consideration band. Scores below 74 can be considered significantly discrepant without further consideration while scores above 84 would not even be eligible for further consideration. Scores falling within the band would require careful consideration to determine the real significance of the discrepancy. I hope this makes more sense. If not, let me know and I will work on the visual display.