The average time per item on each of the three core tests is shown in seconds. As with memory span, the data are also not suitable for standardisation in the conventional sense, because the basic psychometric principles of normality of distribution are not met. In this case, the most obvious violation of those principles is that skewedness exceeded acceptable limits. This is very typical of time-related data, which tends to be distributed with a high positive skew (i.e. scores bunching to the lower or left tail of the distribution, with a long, thin, extended right tail). So instead of presenting memory span in terms of standard scores and centiles, a simple comparison with the age group is provided in three bands: ‘fast’ (less than one standard deviation below the mean time per item of the standardisation sample), ‘average’ (between one SD below and one SD above the mean time per item of the standardisation sample), and ‘slow’ (more than one SD above the mean time per item of the standardisation sample).