Early Intervention In Autistic Spectrum Disorders (Lovaas)

Baseline and Follow-Up Assessment

The following notes, stimulated by discussion at a working group established to advise on a consistent approach to early intervention, set out to highlight what would be good practice in terms of initial assessments of the young children concerned and of gaining baseline data by which to measure change over time.

Some of the difficulties are set out; advice is quoted from professionals with much experience in this field; and suggestions are made for the actual means of assessment.

M.J.Connor     January 1999



* Although all characteristically autistic attributes and behaviours may be part of the 'triad' of disabilities, the ways in which they are expressed and their severity may differ markedly across cases. Therefore, individuals diagnosed as having an ASD will show a complex range of difficulties and styles such that there is a need for wide-ranging assessment approaches to cover all the areas and needs.

* Since skills learnt do not readily generalise from the initial learning situation to any other setting, it may be necessary to ensure that assessments are carried out in a number of places and contexts. Further, observations will assist in highlighting which specific features of the environment are important for the child's demonstration of various skills.

* Specific cues operate in the memory of individuals with ASD so that, for example, a child may not answer a question or attempt a task unless the stimulus is (re-)phrased in such a way as to contain familiar forms of words, or some verbal "trigger". To ensure the best performance from the child, it may be necessary to provide specific cues thus to make explicit what the child is required to do ; and precise assessment styles may vary from one case to

* It is questionable whether (normative) tests of early development or "intelligence" measure anything either precisely or scientifically. Rather, tests should be interpreted with caution. In addition, two different instruments might be used for baseline testing and for follow-up testing and one should avoid assuming that the two sets of results are directly comparable or that the between-test changes are fully reliable.

* As in the case of any test situation, the performance of a child with an ASD will be influenced by:

Any prior experience of a (formal) assessment setting.

Attention and concentration span.

Familiarity with the assessor (and the assessors familiarity with the child's style).

* Any attempts to measure change among a sample of individuals with ASD and, thus, to evaluate the impact of an intervention, will be inhibited by:

Scarcity of suitable test material (especially for before- and after- testing).

Heterogeneity of the sample.

Clear definitions, and contents, of the approach.

Therapist "drift".

Confounding of different approaches.

Changes in personnel involved, with varying training and skills.

* There may well be a gap between skills amenable to formal or systematic assessment and those required for normal day to day functioning. The latter may include social interaction, friendship forming, social communication, conceptual abilities, obsessional or ritualistic behaviours, and mood disturbance.


Areas for Assessment

The initial diagnosis of an ASD will involve an analysis of a number of skill areas and behaviours.

* Intellectual level (IQ).

* Level of language development - expressive and receptive.

* Behavioural maturity - whether the behaviour observed is appropriate for actual age, mental age, and language age. If the behaviour is not appropriate for these levels, one would also examine:

Social skills.



Other behaviours.

* The existence of any medical conditions.

* The existence of any other psycho-social factors.

Areas which appear particularly relevant to the baseline assessment and review assessment of young children (including those for whom an intensive early behavioural intervention is being implemented) include:


Language and communication.


Social skills.

Self help.

Behaviour (including the existence of tantrums ).

The existence of stereotyped, obsessive, and ritualistic behaviours.


Instruments Cited in Published Books and Papers

Cognitive Ability

Wechsler Pre-School and Primary Scale of Intelligence - Revised UK Edition (Wechsler D. 1990 London : The Psychological Corporation).

Wechsler Intelligence Scale for Children - Third UK Edition (Wechsler D. 1992 London: The Psychological Corporation).

Bayley Scales of Infant Development (Bayley N. 1993 London: The Psychological Corporation).

Merrill-Palmer Pre-School Performance Scale (Stutsman R. 1931 NFER-Nelson).

Raven's Progressive Matrices (Raven J. 1954 NFER-Nelson).

Leiter International Performance Scale Battery (Leiter R. 1979 NFER-Nelson).

Differential Ability Scales (Elliott C. 1990 San Antonio : Psychological Corporation).

Kaufman Assessment Battery for Children (Kaufman B. 1983 NFER-Nelson).

McCarthy Scales of Children's Abilities (McCarthy D. 1972 London : The Psychological Corporation).

Stanford-Binet Scales - 4th. Edition (Thorndike R., Hagen E., and Sattler J. 1985 NFER-Nelson).

Cattell Infant Scale [downward extension of the Stanford-B met Scales] (Cattell P. 1960 London : The Psychological Corporation).

Language Development

British Picture Vocabulary Scale (Dunn L., Dunn L., Whetton C., and Burley J. 1997 NFER-Nelson).

Reynell Language Scales (Reynell J. and Huntley L. 1985 Revised by Edwards et al 1997 NFER-Nelson).

Expressive One Word Vocabulary Test (Gardner J. 1990 Los Angeles Western Psychological Services).

Sequenced Inventory of Communication Development (Strain P. and Cordisco L. 1994 LEAP Preschool... In Harris and Handleman (Eds) Preschool Education of Children With Autism. Austin, Texas : Pro-Ed).

Pragmatic Profile of Everyday Communication Skills in Children (Dewart H. and Summers S.1995 NFER-Nelson).


Developmental Play Rating Scale (Dawson G. and Galpert L. 1990 Development and Psychopathology 2 151-162).

Symbolic Play Test (Lowe and Costello 1988 NFER-Nelson).

Behaviour Checklists and Rating Scales

Vineland Adaptive Behaviour Scale (Sparrow S., Balla D., and Cicchetti D. 1984 Circle Pines American Guidance Service).

Childhood Autism Rating Scale - CARS (Schopler E., Reichler R., and Renner B. 1988 Los Angeles : Western Psychological Services).

Autism Behaviour Checklist (Krug D., Arick J., and Almond P. 1980 Journal of Child Psychology and Psychiatry 21 221-229).

Psycho-educational Profile (Schopler E., Reichler R., Bashford A., Lansing M., and Marcus L. 1990 Austin, Texas Pro-Ed).

Autism Diagnostic Interview Revised - ADI/R (Lord C., Rutter M., and LeCouteur A. 1994 Journal of Autism and Developmental Disorders 24 335-349).

Behavioural Rating Instrument for Autistic and Other Atypical Children (Ruttenberg B. Kalish B., Wenar C., and Wolf E. 1977 Philadelphia Developmental Centre for Autistic Children).

Handicaps and Behaviour Schedule - HBS (Wing L. and Gould J. 1978 Journal of Autism and Childhood Schizophrenia 8 79-97).

Autism Diagnostic Observation Schedule (Lord C., Rutter M., Goode S., Heemsbergen J., Jordan H., and Mawhood L. 1989 Journal of Autism and Developmental Disorders 19 185-212).

Autism Diagnostic Observation Schedule [Generic) - ADOSIG (Lord C., Rutter M., and DiLavore P.1996 Unpublished MS., University of Chicago).

Pre-linguistic Autism Diagnostic Observation Schedule - PL/ADOS (DiLavore P., Lord C., and Rutter M. 1995 Journal of Autism and Developmental Disorders 25 355-379).


Choice of Assessment Measures

Advice was sought from the Birmingham team (Jordan et all 1998), who produced the DfEE Research Report No.77 on educational interventions, with the following suggestions:

Cognitive Ability

For young children without good language: Sneiders-Ooman Test or Leiter Test. For assessing further thinking skills : Ravens Matrices.

(The implication is for the use of WPPSI or equivalent.. such as the British Ability Scales [Elliot C., Murray D., and Pearson L. 1996 NFER-Nelson] for children with adequate language)

Language Level

Language comprehension : TROG and BPVS.

Communication : Pragmatic Profile of Everyday Communication Skills in Children.


The general advice is to undertake structured observation of the child's interaction with parents, siblings, known peers, and unknown peers; and to record on videotape the child's performance for further analysis and for comparison with later behaviours. Vineland Scales are recommended if published instruments are to be used.

Autistic Behaviour

Autism Diagnostic Interview or, as a slightly less favoured alternative (because of its close association with TEACCH), Childhood Autism Rating Scale.

Advice was also gained from Professor Howlin and her colleagues at St. George's Hospital Medical School who are embarking upon a research project to evaluate the effectiveness of early behavioural intervention programmes.

Their plans include the use of assessment measures to record (non-verbal) IQ, and receptive and expressive language. Further, all children accepted as part of the project must reach the diagnostic criteria on the Autistic Diagnostic Interview. other measures will be used to record changes over time.

The actual proposals are as follows:

Cognitive Ability

Bayley Scales ; Merrill-Palmer or WPPSI, according to age and ability, (or WISC as appropriate for age).


Reynell Language Scales ; BPVS ; Expressive One Word Vocabulary Test ; TROG.


Symbolic Play Test; Vineland Adaptive Behaviour Scales.

Autistic Behaviour

Autism Diagnostic Interview ; Autism Diagnostic Interview Schedule.

In addition, families will be asked to complete diary records in order to monitor the amount and type of their interaction with the child and to highlight the particular behaviours and skills that have been subject to attention; and baseline measures of parental stress will be gained.

Subsequently, further information will be gathered in respect of school placement, teacher assessment of the child's performance, and follow-up ratings of parental stress.

Current Conclusion

With regard to the extent of overlap between the two sets of advice, the matter of relative ease of access to material, and the wish on the part of the St. George's team to replicate as far as possible the original Lovaas-type studies, it would seem reasonable to follow the pattern set out by Howlin and her associates.

The areas for initial and follow-up assessment would include IQ, diagnostic data, and a

measure of the level of symptoms, interaction, and language.

It is suggested that actual measures would include:

1. A choice from Bayley Scales ; Sneider-Ooman Test; WPPSI ; Merrill-Palmer; WISC;

British Ability Scales.

2. Autistic Diagnostic Interview; Vineland Adaptive Behaviour Scales.

3. Symbolic Play Test.

4. A choice from TROG ; BPVS ; Reynell Scales ; Expressive One Word Vocabulary Test;

Pragmatic Profile.

5. Structured observation of child's interactions in various settings ; and the use of the Parental Involvement Project Developmental Charts (Jeffree D. And McConkey R. 1976 Hodder and Stoughton) may be used to gain further data from the parents.

(6. Possibly.... parental stress questionnaire, such as Parenting Stress Index [Abidin R. 1995 NFER-Nelson)).

It is recognised that the use of IQ tests has been much debated over recent years, but their inclusion in this context is justified on the grounds that....

1. It may be useful in avoiding any confounding of diagnoses (i.e. ASD with severe learning


2. It is in line with existing practice.

3. The tests provide a measure of day to day functioning in a number of areas.

4. The tests provide a clear structure within which to observe the child's general style and behaviour.

The use of the P.I.P. Developmental Charts is suggested for possible inclusion since the charts provide a clear structure and format for the eliciting of parental information, cover a number of areas in some detail, and can be used for baseline assessment and follow-up thus providing an opportunity to observe change.

The choice of tests may be governed, at least in part, by the age of the child (as well as the level of language development ), and by the need to ensure that initial and subsequent testing produce results that are directly comparable.

Some of the assessments, such as TROG or Reynell, would be shared with the speech and language therapist. Assessments may also be carried out by other professionals, such as clinical psychologists or clinical medical officers, and a consistency of approach is to be sought.

M.J.Connor Autism (Lovaas) Working Group January 1999

This article is reproduced by kind permission of the author.

© Mike Connor 1998.

Back to NAS Surrey Branch Welcome Page