The Effect of MAOA and Stress Sensitivity on Crime and Delinquency: A Replication Study

• Author: Jonathan R. Brauer,Scott F. Stoltenberg,Jukka Savolainen,Joseph A. Schwartz,Christa C. Christ
• DOI: 10.1177/1043986218770001
• Published date: 01 August 2018
• Date: 01 August 2018
• Subject Matter: Articles

https://doi.org/10.1177/1043986218770001

Journal of Contemporary Criminal Justice

2018, Vol. 34(3) 336 –353

© The Author(s) 2018

Reprints and permissions:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/1043986218770001

journals.sagepub.com/home/ccj

Article

The Effect of MAOA

and Stress Sensitivity on

Crime and Delinquency:

A Replication Study

Christa C. Christ1,2, Joseph A. Schwartz3,

Scott F. Stoltenberg2, Jonathan R. Brauer4,

and Jukka Savolainen5

Abstract

Across several meta-analyses, MAOA-uVNTR genotype has been associated with

an increased risk for antisocial behavior among males who experienced early

life adversity. Subsequently, early life stress and genetic susceptibility may have

long-term effects on stress sensitivity later in life. In support of this assumption,

a recent study found evidence, in two independent samples, for a three-way

interaction effect (cG × E × E) such that proximate stress was found to moderate

the interactive effect of MAOA-uVNTR and distal stress on crime and delinquency

among males. In light of recent developments in cG × E research, we attempted to

replicate these findings in an independent sample of university students. Our results

failed to support any cG × E or cG × E × E effects reported in the original study.

Implications of a failed replication and general concerns for future cG × E research

are discussed.

Keywords

monoamine oxidase A gene, antisocial, distal stress, proximate stress, childhood

maltreatment

1Hamilton College, Clinton, NY, USA

2University of Nebraska–Lincoln, NE, USA

3University of Nebraska, Omaha, NE, USA

4Indiana University, Bloomington, IN, USA

5University of Michigan, Ann Arbor, MI, USA

Corresponding Author:

Christa C. Christ, Department of Psychology, Hamilton College, Clinton, NY 13323, USA.

Email: cchrist@hamilton.edu

770001CCJXXX10.1177/1043986218770001Journal of Contemporary Criminal JusticeChrist et al.

research-article2018

Christ et al. 337

Introduction

Candidate gene-by-environment (cG × E) research has grown exponentially in recent

years, with applications to a variety of social science topics. Concern for the integrity

of this agenda has also increased due to low overall replicability of scientific literature

more broadly (Baker, 2016), and cG × E findings specifically (Duncan & Keller, 2011;

Risch et al., 2009). A recent article provided recommendations to address the chal-

lenges of cG × E research in response to the issues brought forth by the National

Institute on Alcohol Abuse and Alcoholism (Dick et al., 2015). We find these recom-

mendations, subsequently discussed, to be a compelling means toward efficient prog-

ress in understanding the underlying genetic architecture of complex behavioral

phenotypes, and we rely on them here to evaluate current cG × E research.

The purpose of cG × E research is to examine the potential contribution of specific

genetic variants to an observable outcome (i.e., “phenotype”) in the context of varying

environmental factors. However, because the effect of any one genetic variant is likely

to be extremely small (Chabris, Lee, Cesarini, Benjamin, & Laibson, 2015; Kendler,

2013), large samples (e.g., N > 1,000) are needed to secure sufficient power for detect-

ing a main effect of a candidate gene and, therefore, even larger sample sizes would be

needed to reliably detect cG × E effects (Heo & Leon, 2010). With this in mind, inde-

pendent replications of cG × E findings prior to publication are recommended, but also

do not sufficiently address all the issues raised by Dick et al. (2015).

In addition to a lack of sufficient statistical power, two additional methodological

concerns with cG × E research are noteworthy. First, genetic variants are often selected

and modeled without a clear theoretical link to the phenotype of interest. In particular,

studies targeting candidate genes of the “usual suspect” variety, such as MAOA, are

inconsistent with the polygenic contribution to most complex traits (Dick et al., 2015).

Second, there is considerable heterogeneity in the operationalization and measurement

of environmental factors (Dick et al., 2015). The review of “stress” measurements in

particular has been shown to be widely variable in terms of validity, severity, type, and

time period, suggesting that replication of studies involving stress as an environmental

moderator are essential (Monroe & Reid, 2008; Uher & McGuffin, 2010). Despite

these limitations, the existing literature appears to provide strong a priori reasons to

consider the MAOA-uVNTR genotype as a potential moderator of the association

between exposure to stressful experiences and criminal behavior. The results of mul-

tiple meta-analyses support the finding that MAOA-uVNTR genotype moderates the

association between environmental adversity and risk for antisocial behavior (Byrd &

Manuck, 2014; Ficks & Waldman, 2014). Moreover, the MAOA-uVNTR has a known

effect on the rate of transcription of the MAOA gene (Sabol, Hu, & Hamer, 1998).

Based on these findings, and in light of the methodological issues surrounding

cG × E research, the present study aims to replicate the results from a recent article by

Wells et al. (2017) focusing on the interactive effect of MAOA-uVNTR genotype and

stress on delinquent behavior. The findings from Wells et al. (2017) found proximal

life stress to be associated with increased risk for criminal behavior in males. Critically

important for the present study, they found this association to be stronger among males