Paratuberculosis is a chronic infection, and animals are not equally affected by it. Therefore, diagnostic tests that are able to detect different stages of the infection are needed for objective decision making. A longitudinal study was carried out to describe the ability of 2 tests to predict 2 conditions in dairy cattle: "infection" and "infectious," exemplifying 2 different purposes of testing. "Infection" is the term of choice for certification and eradication purposes, and "infectious" is more relevant for control purposes. In the study period of 3 yr, repeated sampling of milk (n = 23,219) and feces (n = 8,832) was performed. A total of 1,985 Danish dairy cows provided material for the study. Milk samples were analyzed for antibodies using an ELISA, and fecal samples were analyzed for mycobacteria by culture. A reference test to correctly classify cattle antemortem does not exist; thus, "infection" and "infectious" were defined by repeated testing using one test as the condition to be detected by the other test. Fecal culture responses were evaluated against antibody status, and ELISA responses were evaluated against detected bacterial shedding. The results of this study indicate that the ability of both tests to detect "infection" increases almost linearly from 2 to 5 yr of age, whereas the ability of both tests to detect "infectious" is not affected by age. Purpose-specific tests are required to appropriately interpret and use test results for management of paratuberculosis, and relevant covariates, such as age, should be included when possible.