Tier 2 studies would be those using existing samples or data to evaluate an a priori formulated hypothesis, where the biomonitoring this website strategy was not specifically designed for this purpose. In Tier 3 studies, the research relies on existing samples or data without a pre-specified hypothesis or involves multiple simultaneous hypothesis testing. We recognize that at present, the research rationale for most biomonitoring studies involving short-lived chemicals will be described as Tier 3 studies. Evaluative schemes for participant selection apply to studies of both persistent and short-lived

chemicals. The goal of participant selection in epidemiological research is to build a “bridge” between information that is obtainable from the sample and information sought about the target population (Kalsbeek and Heiss, 2000). The actual process

of selecting an unbiased population sample is an ongoing challenge in case–control, longitudinal (cohort) and cross-sectional studies (Vandenbroucke et al., 2007). The issue of participant selection is not unique to epidemiological research of short-lived chemicals. Yet biomonitoring studies may not pay sufficient attention to this problem. Previous reviews of biomonitoring studies presented evidence that selection bias may represent an important threat to internal validity (Bull et al., 2006 and Faust

et al., 2004). The same concerns are also applicable to biomonitoring studies of ALK inhibitor short-lived chemicals such as phthalates (Durmaz et al., 2010, Wang et al., 2013 and Wirth et al., 2008). Tier 1 studies include an unbiased selection and/or follow up protocol with a high (e.g., over 80%) response rate in cross-sectional or case–control studies, or low (e.g., less than 20%) loss to follow up in cohort studies. Tier 2 studies have an unbiased selection/follow up protocol and a low (e.g., 50%–80%) response rate in cross-sectional or case–control studies, or high (e.g., 20%–50%) loss to follow up in cohort studies. Tier 3 studies are those that include less than 50% of eligible participants, or fail to report methods of sample Acyl CoA dehydrogenase selection and/or rates of non-response or loss to follow up. A study that does not report this information should be assumed to be a Tier 3 study. It is important to keep in mind that a low response rate or a high frequency of loss to follow-up should not be equated with selection bias. Selection bias occurs when the proportions of persons included in the final dataset (a.k.a. selection probabilities) differ by both exposure and outcome (e.g., among exposed cases, non-exposed cases, exposed non-cases and non-exposed non-cases.