Twin designs, comparing correlations in monozygotic (MZ) versus dizygotic (DZ) twins, have an extensive history. One major confounder in such studies is that MZ twins may share postnatal environmental influences more so than do DZ twins. To avoid such confounding, twins separated at or soon after birth have been studied, but their scarcity often makes this approach impractical. Another method has been to measure the degree of contact twins have maintained over time, and adjust the observed correlations. Here, we remove confounding by utilizing the discrepancy between biological and self-perceived zygosity to separate environmental from genetic sources of twin similarity. We analyzed dietary patterns and physiologic traits in 350 female twin pairs of the 1988 Kaiser Permanente Twin Registry. Among twin pairs, 175 were MZ by self-report and genetic testing (MZC), 136 were DZ by self-report and genetic testing (DZC), 30 were MZ by genetic testing but not by self-report (MZW), and 9 were DZ by genetic testing but not by self-report (DZW) but were excluded due to small sample size. For healthy food patterns, MZC and MZW intraclass correlations were similar and greater than for DZC, yielding positive and significant heritability estimates. For unhealthy food patterns, the MZC, MZW and DZC correlations were similar with no significant heritability. For physiologic traits, MZC and MZW correlations were similar and higher than those for DZC, indicating significant heritability, except for insulin for which MZW and DZC were similar and which showed modest heritability. Twins of mistaken zygosity (TOMZ) provides a useful approach to robust determination of heritability.