In Reply We thank Levy for the comments on our Research Letter.1 The comment stated that by using classroom-level measures, we ignored within-individual variance, between-individual variance, and variance across age and artificially reduced the confidence intervals (CIs). We disagree. We used a repeated-measures design, measuring each of the 18 classrooms 3 times during baseline and 3 times during intervention. In this design, between-individual variances, including age, cannot contribute to consumption differences between baseline and intervention periods because the same individuals went through both periods. The design allows us to test whether such differences were attributable to the intervention or simply reflect random error within the individual. The design does not artificially reduce CIs: size of CIs depends on standard errors, computed as standard deviation divided by sample size (number), and classroom-level data may have a smaller standard deviation but also a smaller sample size than individual-level data (18 classrooms vs 235 children), and therefore, CIs may not be smaller in classroom-level data compared with individual-level data. Age could moderate the intervention effect, but such effect may not be detectable given the small age variation in our sample (ages 3-5 years; mean [SD] age of 3.8 [0.65] years).
Melnick EM, Li M. Comments Regarding Nutrition Study Among Preschool Children—Reply. JAMA Pediatr. 2019;173(3):293–294. doi:10.1001/jamapediatrics.2018.5041
Coronavirus Resource Center
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: