Stability of the ADI-R
comparable at both time points, total ADI-R scores were computed using only the NVC domain. This
total ADI-R score (i.e. ADI-R total =RSI + NVC + RBSP) was used to examine overall changes in
symptomatology from T1 to FU.
Typically, on the ADI-R the informant provides details both on current symptoms and behaviour at 4-5
years. For the present study information was collected on the children’s current behaviour at both T1
and FU in order to facilitate comparisons across the two time points. Four questions on the ADI-R vary
according to the age of the child (‘play with peers’, ‘friendships, ‘circumscribed interests’ and
‘imaginative play with peers’) and were only administered when chronologically appropriate. Thus,
they were not administered at both T1 and FU for all individuals and to ensure comparability across
assessments these items were not included in the item level comparisons.
Cognitive ability: Four assessment tools were employed: the Merrill-Palmer Scale of Mental Tests
(MP; Stutsman, 1948), the Bayley Scales of Infant Development (Bayley, 1993), the Wechsler Pre-
school and Primary Scale of Intelligence (WPPSI; Revised and III version; Wechsler, 1990; 2003) and
the Wechsler Intelligence Scale for Children (WISC-IV; Wechsler, 2004). Due to changes in age and
ability between T1 and FU, it was not always appropriate to use the same test at both time points,
although every effort was made to use the same instrument over time as far as possible.
Selection of the appropriate assessment was based on chronological age, developmental level and
language ability. To avoid repeated, separate comparisons of T1 and FU cognitive scores obtained from
different tests and in order to utilise the most relevant IQ data, a “best test” IQ and Mental Age (MA)
score was created for each child using the most age appropriate/ best standardised test available at each
time point according to the following hierarchy: WISC > WPPSI (higher level) > Bayley> MP>