Are Children Performing Better on the Digit Span Backward Task Than the Digit Span Forward Task? An Exploratory Analysis.

Abstract

This research note was motivated by community speech-language pathologists (SLPs) who have expressed concerns about how to interpret the unexpected observation of higher performance on the Digit Span Backward task than the Digit Span Forward task on the Test of Integrated Language and Literacy Skills (TILLS). We therefore conducted an exploratory analysis to examine the pattern of digit span performance in children to help clinicians interpret their findings. We investigated performance on the Digit Span Forward and Digit Span Backward subtests using the TILLS normative sample (N = 1,262). Raw scores, standard scores, and percentile ranks were systematically compared. In addition, we were able to estimate span length (longest number of items that can be accurately recalled) for 69 participants and evaluated the relation between performance and span length. Our exploratory analyses revealed that "better" performance on the backward than forward task was rare. Instead, the pattern that SLPs are reporting can be explained by three factors: (a) a statistical phenomenon exacerbating small differences, (b) comparing standard scores or percentile ranks instead of raw scores or span length, and (c) better backward scores can still mean longer forward span length. It is likely that community SLPs tended to report cases of better backward than forward recall because it is the more remarkable and interesting finding, leading to the perception that the pattern is common, when it is not. Overall, we caution clinicians against overinterpreting their own client's performance on the digit span tests.