Randomised clinical trial of brief behaviourally-oriented interventions for pre-delinquents

Criminology Research Council grant ; (10/85)

This study sought to:

  • design and standardise a brief teacher-rated behaviour scale suitable for assessment of conduct disorders in Australian pupils aged 10 to 15 years; and
  • evaluate a short-term prevention-oriented cognitive- behavioural intervention which focuses on conduct disordered pupils.

Development of the scale took place in two stages. In the first stage, the reliability of the scale items was ascertained by (1) testing whether the responses of two different teachers completing the scale independently about the same child agreed; and (2) testing whether the response of one teacher completing the scale about a particular child at two different times agreed. Adequate reliability was demonstrated for 26 of the 35 scale items. In the second stage, the construct validity (the extent to which the test may be said to measure a theoretical construct or trait) of the 26-item scale was tested by factor analysis. This method identified five factors (acting out, distractability, peer relations, immaturity/withdrawal, and delinquency) which identify conduct disorders.

It was concluded that the 26-item scale has several advantages including its suitability for an Australian sample. The researchers recommend that the scale form part of a battery of tests to be used to accurately measure behaviour problems.

The evaluation study comprised a randomised clinical trial with one and six-month follow-up. The sample included 61 experimental pupils and 62 control pupils. Data were collected from three sources: parents and guardians; pupils; and teachers. While some non-significant improvements were noted, the results failed to indicate global improvements, maintenance and generalisation. First, parent measures failed to show a statistically significant improvement. However, non-significant improvements were noted in social competence and negative behaviour. Similar results were found for pupil measures; however, a non-significant improvement in pupils' status scores was noted. Teacher measures also failed to indicate a statistically significant improvement in experimental pupils. Instead, a significant deterioration was noted in the short term in general behaviour (as measured by the Achenbach Teacher Report Form total behaviour score). In addition, a non-significant improvement was noted in the short-term in more specific conduct (as measured by the Endeavour Behaviour Rating Scale total behaviour score). These results suggest a number of possible recommendations for future interventions.

These include ensuring that selection criteria are met; incorporating environmental support (e.g. from parents and teachers); lengthening the intervention period; and incorporating more intensive programming for generalisation.