Zygosity determination is generally carried out by different methods in small or large twin samples. The probability method based on sex and genetic markers is limited to relatively small samples, as a consequence of its cost. The empirical questionnaire method is applied in several large twin registers. Its margin of error is low enough for population studies, its cost is negligible, but its accuracy is insufficient when zygosity of twin pairs included in definite samples must be individually assessed. Efforts to bridge the distance between the two methods should be made, and they may take either direction: (1) find new, inexpensive genetic markers, or (2) increase the number and accuracy of empirical methods. The accuracy of a number of empirical methods applied to a twin sample of established zygosity has been compared. One modification of the “two peas in a pod” method, originally called “equivocalness method”, appears to warrant inclusion in questionnaire methodology. Also, compound probability as expressed by several empirical methods may reach an acceptable level of accuracy in zygosity assessment.