Published online by Cambridge University Press: 21 February 2012
Tishler and Carey (this issue, 2007) have analyzed the difference in ever smoking prevalence between monozygotic (MZ) and dizygotic (DZ) twins in 8 published twin studies. They conclude that the lower prevalence in MZ twins might invalidate conclusions from classic twin models. I critique the conducted meta-analysis and suggest additional analyses to be conducted. A reanalysis of the Finnish Twin Cohort 1975 questionnaire data (included in Tisher and Carey) indicates that when intrapair correlations are correctly taken into account, the MZ-DZ difference in that study is no longer statistically significant. Even if the Tisher/Carey finding holds up in further analyses, it pertains only to the observed trait, the effect is small in size, and the mechanism underlying it needs to be explored.