Although many radiation epidemiological studies have been carried out, there is still uncertainty about the health effects of low dose and low dose-rate radiation in humans. One reason for this uncertainty is that the risk of radiation itself may be too small to detect. Another possible reason is that the main components of cohorts or statistical method vary in each study. Comparing the Excess Relative Risks (ERRs) with other studies is often one approach; however, few studies have denoted the validity of comparing ERRs. To verify the differences in study methods, we summarized them and the results of radiation epidemiological studies to date. Some of these studies targeted high background residents or patients who received CT scans. In the present work, we focused on cohort studies among nuclear industry workers because they assured more accurate dose measurements and had no possibility of reverse causation (i.e., patients who received CT scans had worse health conditions, which prompted the need for the scans). In addition, we limited the studies to those that summarize derived excess relative risks of mortality based on a linear model.