Figure 1.
Figure 1.

Examples of common outcome variables with normal (ADG; A), binomial or skewed (mortality risk; B), and discrete (clinical illness scores; C) data distributions.

 


Figure 2.
Figure 2.

Diagnostic test illustration of the influence of the commonness of a characteristic (being persistently infected [PI] with bovine viral diarrhea virus [BVDv]) on the interpretation of test outcomes when the test characteristics (diagnostic sensitivity and specificity) remain constant. The interpretation of a positive test result (positive predictive value) is very different when the prevalence of the tested condition is low (A) compared with when the prevalence of the tested condition is high (B), even though the test characteristics do not change.

 


Figure 3.
Figure 3.

Depiction of studies (n = 100) to evaluate the expected results of discovery (A; 10% of trials have treatment or observation groups that are truly different) or confirmatory studies (B; 40% of trials have treatment or observation groups that are truly different) where trials identifying treatment or observation group differences are denoted with an “X” and the true state of the natural world is denoted by no difference between observation groups or treatments