Understanding and Preventing Violence, Volume 2: Biobehavioral Influences (1994)
Chapter: Danish Twin Study
characterized by lower education, lower occupational status, and increased rates of alcohol problems.
With a broad definition of criminal involvement (legally punishable behavior reported to the criminal register), there was little difference in concordances—37 and 31 percent for monozygotic (MZ) and dizygotic (DZ) pairs. Slightly better discrimination was found with a stricter definition of criminal involvement (crimes of violence, sexual assault, theft and robbery)—41 and 26 percent, respectively, for MZ and DZ twins—but the difference does not reach significance. Unfortunately, base rates for criminal registration were not provided. Neither were analyses performed jointly on alcohol use and crime. An intriguing but statistically nonsignificant finding that emerged through the interviews was that concordant identical twins tended to collude more often in the same criminal act than fraternal pairs.
Danish Twin Study
Christiansen (1968) identified all twins born on the Danish islands between 1880 and 1910, where both members survived to age 15, and traced the twins through national and local police or penal registers. Updates of the sample were provided by Christiansen (1974, 1977), Cloninger et al. (1978), and after Christiansen's death, by Gottesman et al. (1983) and most recently by Cloninger and Gottesman (1987).
On virtually every type of classification for crime, MZ twins are more concordant than DZ twins. One illuminating finding is that despite base rate differences in criminal registration between males and females, tetrachoric correlations of liability are remarkably similar—.74 for both male and female MZ twins, .47 for male-male DZ and .46 for female-female DZ pairs. In contrast, the correlation for male-female pairs is .23, suggesting either sex-by-genotype interaction (i.e., the loci that contribute to individual differences in males do not have identical effects in females) or common environment-by-sex interaction (i.e., those environmental factors that contribute to twin concordance do not have the same effects in males and females).
One major difficulty in interpreting these twin results is the difference in base rates for the three zygosity groups illustrated in Table 5. In both genders, MZ twins have the highest—and opposite-sex DZ twins the lowest—registration rates. The difference is significant for males and almost reaches significance for females. Such a pattern suggests an "imitation" or collusion effect
