In a report from the laboratory, the authors learned that values for serum concentrations of cotinine (a biomarker for exposure to passive smoking), which were below the assay's detection limit had registered as missing data; that is, some values of the main exposure were so low that they were undetectable by the assay and left out of your analysis (i.e., did not enter your 2 x 2 table). When the authors went back to their analysis, they noted that participants without a TB diagnosis were more likely to have cotinine values that were below the limit of detection.

Also What process is this describing?

a. Selection dependent on exposure and disease

b. Selection independent of exposure but dependent on disease

c. Selection dependent on exposure but independent of disease

d. Non-differential misclassification of exposure

e. Differential misclassification of exposure by disease status