The intersection of individual differences, personality variation, & military service: A twin comparison design

ABSTRACT

In societies where military service is voluntary multiple factors are likely to affect the decision to enlist. Past research has produced evidence that a handful of personality and social factors seem to predict service in the military. However, recent quantitative genetic research has illustrated that enlistment in the military appears to be partially heritable and thus past research is potentially subject to genetic confounding. To assess the extent to which genetic confounding exists, the current study examined a wide range of individual-level factors using a subsample of twins (n = 1,232) from the restricted-use version of the National Longitudinal Study of Adolescent to Adult Health. The results of a series of longitudinal twin comparison models, which control for the latent sources of influence that cluster within families (i.e., shared genetic and family factors), illustrated generally null findings. However, individuals with higher scores on measures of extraversion and the general factor of personality were more likely to enlist in the military, after correction for familial confounding. Nonetheless, the overall results suggest that familial confounding should be a methodological concern in this area of research, and future work is encouraged to employ genetically informed methodologies in assessments of predictors of military enlistment.

KEYWORDS:

• Military enlistment
• genetic confounding
• twin comparison
• Add Health

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can accessed https://doi.org/10.1080/08995605.2020.1786323.

Notes

1. All participants in the Add Health study provided written informed consent for participation in all aspects of the study (https://www.cpc.unc.edu/projects/addhealth/faqs/addhealth/index.html#Was-informed-consent-required). Given that the data employed in the current study are secondary and de-identified, the Institutional Review Board at the University of Cincinnati determined that the data did not meet the regulatory criteria for research involving human subjects.

2. Models were reassessed using multiple racial category variables (i.e., White [reference], Black, American Indian, Asian, Other, and Hispanic ethnicity) as covariates and the results were virtually identical to those reported herein and no differences in terms of the likelihood of enlistment were observed across the different racial categories. Thus, for the sake of parsimony the dichotomous measure of race is employed in the current study.

3. The between-family and within-family variables were created for each covariate following a two-step procedure. First, an overall family mean for each independent variable was created by averaging the scores for two twins from the same family on the corresponding independent variable. Second, a mean deviation score for each twin within a twin pair was created by subtracting a twin’s score on an independent variable from the family mean for the same variable.

4. For reference, the mean age at Wave 4 was 29.05 years (SD = 1.64, Min., Max. = 25, 33).

5. We note that the p-value associated with this coefficient is precisely.05 so it is considered statistically non-significant. Nonetheless, given the stringent nature of the modeling strategy and the breadth of the associated 95%CI we highlight this effect herein.