To investigate the effect of observing school meals on children's dietary reports.

One hundred and twenty children randomly selected, but with half girls, from usual school-meal eaters among 312 volunteers (from all 443 fourth-grade children in six schools in one district).

Children were assigned randomly to one of 12 conditions yielded by crossing observation status (observed; not observed), target period (previous day; prior 24 hours), and interview time (morning; afternoon; evening).

Response variables included interview length, number of meals and snacks reported for the target period, and, for two school meals, number of meal components reported, importance-weighted number of items reported and kilocalories reported. These variables were transformed to principal components; two were retained (1, the school meal variables; 2, interview length and number of meals and snacks). Analyses of variance on principal component scores tested effects of observation status, target period, interview time and all interactions. Observation status did not affect scores on either retained principal component. Scores on Component 2 showed that more intake was reported in prior-24-hours interviews than in previous-day interviews.

The effect of target period on reported intake indicates that the response variables were sufficiently reliable to detect manipulations. This, together with the finding that response variables did not depend on observation status, suggests that observation of school meals does not affect fourth-grade children's dietary reports, and that conclusions about dietary reports by fourth-grade children observed eating school meals in validation studies may be generalised to dietary reports by comparable children not observed.