Stark 1996.

Study characteristics
Methods	Parallel RCT with half participants receiving intervention first then other half 3 months later ‐ not reported. 4 families changed group after randomisation ‐ due to conflicting vacation scheduling. Thus not truly randomised.
Participants	N = 10. 1 withdrew from control group after randomisation. Total sample n = 9. Age range: 5.3 years to 10.1 years. Mean age: 7.3 years (SD = 1.7).
Interventions	1. Group behavioural intervention (n = 5). 7 weekly sessions ‐ baseline assessment plus snack, breakfast, relaxation skills training, lunch, dinner and maintenance strategies targeted over following 7 sessions. 2. Wait list control (n = 4). Parent meeting and 7‐day food diaries at times corresponding to baseline and last week of intervention.
Outcomes	1. Calorie intake. 2. Anthropometric measures ‐ weight, height and skinfold. 3. PFT. 4. REE. 5. Physical activity ‐ Caltrac electronic accelerator. Assessments at baseline, then 3 months and 6 months post‐treatment.
Notes	Unclear what population the sample was drawn from. POI: Daily calorie intake. ITT ‐ unclear. Blinding ‐ unclear. No significant difference on z scores for weight or age before treatment. Control group average SES group II ‐ technical workers. Intervention group average SES III ‐ skilled labour. Equaivalence at baseline doubtful.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The authors did not describe details of random generation process. It is just stated that 'the nine subjects were randomly assigned to either a behavioral intervention or a wait list control group' (Stark 1996).
Allocation concealment (selection bias)	Unclear risk	The authors did not provide information about adequate concealment of allocation.
Blinding (performance bias and detection bias) All outcomes	High risk	Participants and personnel providing the intervention were not able to be blinded due to the nature of the intervention and the study design (wait‐list‐control design). The main outcome 'calorie intake' is assessed via self‐reported daily diet dairies. Due to that the measure is not objective and might be influenced by the knowledge of allocation to one of the conditions. All other objective measures (e.g. weight) are not likely to be influenced by the lack of blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	The authors reported that there was no attrition.
Selective reporting (reporting bias)	Unclear risk	The authors reported alls pre‐specified outcomes. It is unclear if additional outcomes were pre‐specified in the study protocol but not reported.