The GQ and individual scales from the Griffiths Mental Development Scales (2–8 years) have differing means and standard deviations. They differ from the common standard among developmental tests. Direct comparisons between Griffiths subscale scores and between the Griffiths and other test scores are therefore prone to misinterpretation. A corrective formula for Griffiths scores is supplied with tables for popularly used ranges.
- Griffiths Scale
- mental development
- standard score
Statistics from Altmetric.com
The Griffiths Mental Development Scales have a long and developing history in clinical practice and research. The scales show continuing validity over time and across cultures.1 The scales have been used, alongside other measures, in many studies reported by this journal. A feature of the scales' original development was the use of standard deviations for each scale based on a simple ratio transformation of the raw data2: the mental age divided by the chronological age. As table 1 shows, this yielded similar, but slightly different means and standard deviations for each scale, including the aggregate general quotient (GQ).
Comparing scale scores with each other is therefore prone to potential error, if they were considered equivalent. In many cases, especially where the scores are close to the mean, this error is minimal. However, the error is exaggerated as scores move away from the mean. This may lead to errors in interpreting Griffiths quotients, particularly where numerical cut offs for popular developmental and ability ranges are applied to unadjusted scores.
Two solutions are suggested to ensure an informed comparison of scores between individual Griffiths scales and scores from different tests. The first alternative is to calculate an exact standard score equivalent (with a mean of 100 and SD 15). This is the most precise alternative.
Conversion of the Griffiths scale scores into standard scores can be carried out using a simple transformation algorithm,where SS2 is the new adjusted standard score (M 100, SD 15), M2 is the mean of the normal standard score (100), M1 is the mean for the Griffiths subscale or GQ, SS1 is the Griffiths standard score, SD1 is the standard deviation for the Griffiths' subscale or GQ and SD2 is the normal standard deviation (15).
The second option is to develop descriptive category ranges for Griffiths scores that are equivalent to those used by other tests. Table 2 uses standard score ranges (± 2 SD) that are based on those used by popular tests with a mean of 100 and SD 15, for example, the Wechsler Preschool and Primary Scale of Intelligence—revised,3 and the British Ability Scales II.4 Descriptive terms for the score ranges, such as “borderline” or “low” for scale scores of 70–79, and “normal” or “average” for scaled scores of 90–109, refer implicitly to this shared standard. The same score ranges are used by the Bayley Scales of Infant Development.5
Table 3 shows the downward extension of these ranges, with the Griffiths scale score bands for the ICD106 mental retardation ranges. The table highlights the difference between Griffiths subscale scores at the extreme ranges.
These tables and transformations are suggested as simple corrections to potential interpretation error based on the traditionally derived Griffiths standard scores.