Purpose: To examine whether there are differences between demographic, behavioral, and biological variables for those with invalid accelerometry data (IAD) and those with valid accelerometry data (VAD).

Design: Cross-sectional.

Setting: Data from 2003–2006 National Health and Nutrition Examination Survey (NHANES) were used.

Subjects: Participants included 1,315 children (i.e., 6–11 years) with VAD and 534 children with IAD and 1,859 adolescents (i.e., 12–17 years) with VAD and 1,057 with IAD.

Measures: Physical activity (PA) was measured using an accelerometer, with questionnaires used to assess demographic and behavioral variables and biological parameters assessed from a blood sample.

Analysis: Wald and design-based likelihood ratio tests and logistic regression were used to assess differences between those subjects with IAD and those with VAD.

Results: After adjustments, overweight children, compared to normal weight children, were 1.6 (odds ratio [OR] = 1.67; 95% confidence interval [CI]: 1.22–2.29) times more likely to have IAD. After adjustments, and as an example, adolescents engaging in 4 or more hours of computer use per day, compared to no computer use, were 2.6 (OR = 2.6; 95% CI: 1.38–5.18) times more likely to have IAD.

Conclusion: Excluding youth with IAD may introduce bias, limit generalizability, and ultimately underestimate the association between PA and health outcomes. Future research is needed to identify reasons for poor monitoring compliance.