- While a low degree of reproducibility indicates that the questionnaire does not provide a valid measure of long term physical activity (energy expenditure), a high degree of reproducibility does not necessarily ensure validity.

- High reproducibility can simply reflect correlated error (i.e., a misinterpretation of a question on each questionnaire administration) (Figure 2).

- In summary, because reproducibility studies are usually quick and inexpensive to conduct, they are an appropriate part of the questionnaire evaluation, but cannot substitute for validity studies.