Introduction: Considering the imprecise characteristic of the currently used exposure classification categories discussed in the first part of this series article, in this second part, the authors will conduct a series of meta-analyses and subgroup analyses using various exposure classification categories from crude ones to more precise ones. Methods: The medical librarian searched MEDLINE (PubMed), EMBASE, and the Cochrane Library until 16 December 2020. Results and Discussion: As for the separate analysis stratified by each tumor type, the odds ratio for meningioma for regular users was statistically significantly decreased, 0.86 (95% CI 0.77-0.95). For ipsilateral users, the pooled odds ratio for meningioma (1.20 (95% CI 1.04-1.39)), glioma (1.45 (95% CI 1.16-1.82)), and malignant tumors (1.93 (95% CI 1.55-2.39)) showed a statistically significant increased estimate. For years of use, the pooled odds ratio for glioma for > 10 years of use group showed a statistically significant increased estimate (1.32 (95% CI 1.01-1.71)). To sum up, these results stratified by each tumor type, glioma, and malignant tumors were strong candidate tumor types of the possible tumorigenic effect of RF-EMR from wireless phones. In the meta-analysis for the total cumulative hours of use > 867 hours, the pooled odds ratio was 1.56 (95% CI 1.27-1.91). This statistically significant increased pooled odds ratio indicates that a rather moderate amount of cumulative exposure to RF-EMR from wireless phones could induce brain tumors. Conclusion: The study hypothesis, ‘the causal relationship between RF-EMR exposure and brain tumor incidence can be clearly validated only when accurate exposure assessment methods are applied,’ was validated by the increasing pattern of estimated odds ratios from the use of rather crude exposure categories to more precise exposure categories.