Adequate assessment of energy balance-related behaviours in adolescents is essential to develop and evaluate effective obesity prevention programmes. The present study examined the test–retest reliability and construct validity of a questionnaire assessing energy balance-related behaviours in adolescents during the evaluation of the DOiT (Dutch Obesity Intervention in Teenagers) intervention.

To assess test–retest reliability, adolescents filled in the questionnaire twice (n 111). To assess construct validity, the results from the first test were compared with data collected in a personal cognitive interview (n 20, independent from the reliability study). For both reliability and validity, intraclass correlation coefficients for continuous data or Cohen's kappa coefficients for categorical data were calculated as well as percentage agreement.

Data were collected during school time from February to May 2010.

Study participants were Dutch adolescents aged 12–14 years attending pre-vocational secondary schools.

In more than three-quarters of the ninety-five questionnaire items the test–retest reliability appeared to be good to excellent. Moderate reliability was found for all other twenty-one items. Fifty-one items (of ninety-five items) showed good to excellent construct validity. Construct validity appeared moderate in twenty-three items and poor in twenty-one items. Most items with poor construct validity concerned consumption of sugar-containing beverages and high-energy snacks/sweets.

Our study showed good test–retest reliability and largely moderate to good construct validity for the majority of items of the DOiT questionnaire. Items with poor construct validity (most of them found for items concerning energy intake-related behaviours) should be revised and tested again to improve the questionnaire for future use.