Content uploaded by Anthony Hoskin
Author content
All content in this area was uploaded by Anthony Hoskin on Nov 20, 2017
Content may be subject to copyright.
FETAL TESTOSTERONE AND CRIMINALITY: TEST
OF EVOLUTIONARY NEUROANDROGENIC
THEORY∗
Anthony W. Hoskin1and Lee Ellis2
1Criminology Program, University of Texas of the Permian Basin
2Department of Anthropology and Sociology, University of Malaya
KEYWORDS: 2D:4D digit ratio, prenatal testosterone, sex, crime, delinquency, evolu-
tionary neuroandrogenic theory
Evolutionary neuroandrogenic (ENA) theory asserts that criminality is a crude form
of competitive behavior over resources, status, and mating opportunities. Theoreti-
cally, males have been selected for resource acquisitiveness as a result of female prefer-
ences for mates who are successful at resource provisioning. ENA theory also asserts
that brain exposure to both prenatal and postpubertal androgens (particularly testos-
terone) promotes all forms of competitiveness, including those that victimize others.
The present study was undertaken to test ENA theory by correlating 14 self-reported
measures of offending with a biomarker for fetal testosterone exposure based on the
ratio of the 2nd and 4th digits of the right hand (r2D:4D), in a nonrepresentative sam-
ple of 445. Both Spearman correlations and negative binomial regressions produced
results that largely supported the hypothesized connection between offending and high
prenatal androgen exposure, even when findings were analyzed separately by sex. Also,
offending was significantly associated with r2D:4D for both males and females. Over-
all, this study supports the view that exposing the brain to high levels of testosterone
and other androgens prenatally elevates the probability of offending later in life.
Although the biosocial perspective remains a minority approach to criminology
(Cooper, Walsh, and Ellis, 2010), many biosocial theories have been offered in recent
years (Wright, 2010; Wright and Cullen, 2012). Some are Darwinian in nature, arguing
that various forms of criminality have evolved by natural selection—that is, sexual as-
sault (McKibbin et al., 2008; Thornhill and Palmer, 2000), homicide (Buss, 2006; Daly and
Wilson, 1988), child abuse (Jones, 1996; Lightcap, Kurland, and Burgess, 1982), genocide
(Ghiglieri, 2000), and terrorism (Thayer and Hudson, 2010). Other biosocial theories fo-
cus on neurology or neurochemistry, arguing that one or more brain functioning patterns
of offenders differ on average from what is typical of nonoffenders (DeLisi, 2011; Raine,
1997; Walsh and Beaver, 2009; Yang and Raine, 2009).
∗Additional supporting information can be found in the listing for this article in the Wiley Online
Library at http://onlinelibrary.wiley.com/doi/10.1111/crim.2015.53.issue-1/issuetoc.
Direct correspondence to Anthony W. Hoskin, Criminology Program, University of Texas of the
Permian Basin, 4901 E. University Boulevard, Odessa, TX 79762 (e-mail: hoskin a@utpb.edu).
C2014 American Society of Criminology doi: 10.1111/1745-9125.12056
CRIMINOLOGY Volume 53 Number 1 1–20 2014 1
2 HOSKIN & ELLIS
EVOLUTIONARY NEUROANDROGENIC THEORY
A decade ago, a biosocial theory was proposed that contained both evolutionary and
neurological elements (Ellis, 2003, 2005). The theory, called evolutionary neuroandro-
genic (ENA) theory, depicts most criminality as being an expression of crude forms of
competition for resources, status, and mating opportunities. As to why involvement in
crime normally declines as individuals age beyond adolescence and young adulthood,
ENA theory asserts that most people learn quickly that lawful (sophisticated) forms of
competition are usually more effective in acquiring resources, status, and mating op-
portunities (and risk fewer ill consequences) than unlawful (crude) forms. Thus, indi-
viduals who have the greatest difficulties learning and engaging in long-term planning
are those who take the longest time to transition from crude to sophisticated forms of
competition.
At the heart of ENA theory are two propositions: First, males have been favored by
natural selection for being unusually competitive regarding the acquisition of resources,
status, and mating opportunities, which in turn results in their victimizing others more
than females do. Second, androgens (male sex hormones) have evolved in ways that
modify brain functioning to make individuals more competitive and likely to victimize
others.
Beyond these two fundamental assertions, the theory becomes fairly complex in its
explanations for how androgens influence brain functioning to promote competitive
and victimizing behavior. Most of the pertinent variables in this regard are provided in
figure 1 (Ellis, 2005, 2011). This diagram asserts that a wide array of androgen-promoted
neurological, cognitive, learning, and emotional factors interact to impact the probability
of criminal and associated behavior patterns.
According to figure 1, androgens alter three brain functioning patterns during fetal
development. However, these patterns tend not to be fully activated until the onset of
puberty. These three patterns interact with other brain centers affecting foresight and
planning ability (shown in the lower portion of figure 1). According to ENA theory, these
neurological patterns function in ways that shape each individual’s probability of engaging
in delinquency and crime.
The labyrinth of causal connections hypothesized in figure 1 have a limited bearing on
the present study. In fact, only two cells shown in figure 1—those that are shaded—are
directly pertinent to the hypothesis that the current study will test. The shaded cell in the
upper left of the diagram identifies high exposure to perinatal androgens as a key indepen-
dent variable, and the shaded cell in the middle right-hand column specifies criminality
as the dependent variable. In other words, although ENA theory argues that the connec-
tion between prenatal androgens and criminality involves a complex set of intervening
variables, it still unmistakably predicts that exposure to prenatal androgens is central to
elevating the probability of criminal behavior later in life. This hypothesis will be tested
in this study.
MEASURING PRENATAL ANDROGEN EXPOSURE
To test the hypothesis that prenatal androgens promote criminality, one must have
some measure of the presence of androgens during fetal development. Because ENA the-
ory asserts that prenatal and postpubertal androgens both contribute to criminal behavior
FETAL TESTOSTERONE AND CRIMINALITY 3
Figure 1. Graphic Representation of ENA Theory of Criminal and Antisocial Behavior (Excluding the
Evolutionary Portion)
↑Seizuring Susceptibility
↑Suboptimal Arousal
Tendencies
↑Rightward Shift in
Hemispheric Functioning
↑Acting Out, Emotional
Volatility
↑Genes and Prenatal
Environments Promoting
High Prenatal and High
Postpubertal Brain
Exposure to Androgens
Neurological Factors Genetic and Prenatal Factors Cognitive, Learning and Emotional
Factors
Behavioral and Personality Factors
↓Anxiety, Pain Sensitivity
↑Susceptibility to Boredom
↓Attending to Linguistic
and Other Social Stimuli
↓Prosocial Emotions
↓Long-Term Planning and
Coordination of Activities Around
Specific Goals
↓School Performance,
Dependable Work Performance on
Tasks Assigned by Others, and
Success in Business Ventures
↑Victimizing Forms of
Resource/Status Competition
(Criminal and Antisocial Behavior)
↑Externalizing Behavior,
Hyperactivity, Oppositional
Defiance, and Autism
↑Epilepsy, Tourette’s Syndrome,
and Other Seizure Disorders
↓Enlargement and
Efficient Functioning of
the Neocortex
↓Genes and Prenatal
Environments that
Promote Brain
Development, Especially
of the Neocortex ↓Enlargement and
Efficient Functioning of
the Prefrontal Area of the
Frontal Lobes
↓Intelligence and Learning
Abilities
↓Efficient Executive
Functioning
↓Efficient Monitoring and
Regulation of the Limbic
System by Frontal Lobes
↑Competitiveness, Sensation
Seeking, Risk-Taking, Self-
Confidence, and Fearlessness, and
Entrepreneurial Motivations
NOTE: The only association examined in the current study was between prenatal androgen exposure (shown in the shaded cell to the left) and criminal behavior
(shown in the shaded cell to the right).
Source: Ellis 2005: 296.
4 HOSKIN & ELLIS
(as shown in figure 1), one might simply infer prenatal androgen exposure by measuring
postpubertal levels. In this regard, numerous studies have revealed significant positive
correlations between offending and postpubertal levels of testosterone, although some
failures to replicate have been reported (reviewed by Ellis, Beaver, and Wright, 2009:
208–10). However, this is not a very good test of the hypothesis that prenatal androgens
promote criminality because research has shown that there is actually little-to-no correla-
tion between prenatal androgen levels and postpubertal levels, at least when comparisons
are made within each sex (H ¨
onekopp et al., 2007).
Unfortunately, direct measurement of prenatal androgens is difficult. Concentrations
of testosterone in the amniotic fluid can be measured, but doing so carries some risks to
fetuses and is done only with medical justification (van de Beek et al., 2004). Levels of
androgens can be assessed at birth from umbilical cord blood, but these levels seem to
be only modestly indicative of fetal exposure (Mori and Shiota, 1994; van de Beek et al.,
2004). Furthermore, carrying out a study in which umbilical cord blood androgens were
correlated with involvement in delinquency and crime would require following a sizable
sample of individuals for two or more decades.
In the late 1990s, research began pointing toward a surprising indirect measure of pre-
natal androgen exposure. It involved observing a physical trait known as the 2D:4D finger
length ratio (H ¨
onekopp et al., 2007; Manning et al., 1998). As a rule, people with the low-
est finger ratios seem to have been exposed to the highest levels of prenatal androgens.
The term—abbreviated as 2D:4D—refers to the relative lengths of the second and fourth
digits of the hand, with several studies indicating that the right hand provides a more reli-
able indicator of prenatal androgen exposure than does the left hand (Brown et al., 2002;
Grimbos et al., 2010; H ¨
onekopp and Watson, 2010; Kastlunger et al., 2010; Williams,
Greenhalgh, and Manning, 2003). Parenthetically, a study of rats in which testosterone
was injected into some mothers while pregnant but not in control mothers suggests that
the right paw provided a more reliable measure of prenatal androgen exposure than did
the left paw (Talarovicov´
a, Krskov´
a, and Blazekov´
a, 2009). Throughout the remainder of
this report, we will refer to the 2D:4D ratio on the right hand as r2D:4D because most
studies, including the current study, are based on this hand.
It needs to be emphasized that despite its value in estimating prenatal androgen expo-
sure, r2D:4D is far from a perfect indicator (Hell and P¨
aßler, 2011; Lippa, 2003; Manning,
2008). Nonetheless, the few studies that have used more direct methods for assessing pre-
natal androgen exposure have found r2D:4D to be a modest correlate of this exposure
(although estradiol, a biochemical postcursor of testosterone, also seems to affect the fin-
ger length ratio) (Lutchmaya et al., 2004; McIntyre et al., 2005; van Anders, Vernon, and
Wilbur, 2006). Another clear indication that a low r2D:4D ratio is associated with height-
ened prenatal exposure to androgens comes from finding that males have lower r2D:4D
ratios than do females (H ¨
onekopp and Watson, 2010; McIntyre, 2006; van Anders,
Vernon, and Wilbur, 2006).
As to why such a biomarker as 2D:4D would exist, a recent study of mice provided at
least a partial answer. It revealed that the fourth digit on the right forepaw is unusually
dense with androgen receptors. When androgen receptors in newborn male pups were
experimentally inhibited with an anti-androgen drug, the fourth digit grew less relative to
the second digit, thus, producing a higher (i.e., feminized) 2D:4D ratio (Zheng and Cohn,
2011).
FETAL TESTOSTERONE AND CRIMINALITY 5
EVIDENCE THAT PRENATAL ANDROGEN EXPOSURE
PROMOTES CRIMINALITY
Empirical evidence bearing on the possibility that prenatal androgens are associated
with 2D:4D as ENA theory hypothesizes is currently limited to four studies. As we
will summarize, three of the four studies have provided findings supporting theoretical
expectations.
Based on a sample of 77 males, one study reported that males with the lowest r2D:4D
finger ratios had more moving traffic violations than did males as a whole (Schwerdtfeger,
Heims, and Heer, 2010). Another study compared the r2D:4D ratios of 44 males who
have at least one criminal conviction with 46 males who have no convictions. It concluded
that the convicted males had significantly lower r2D:4D ratios (Hanoch, Gummerum, and
Rolison, 2012). However, a study comparing the r2D:4D ratios of 44 male convicts with
66 males who did not have a conviction history found no significant differences between
these two groups (Anderson, 2012). Each of these three studies was based on samples of
100 or fewer individuals, all of whom were males.
The only study conducted so far based on a large sample of both males and females
was one conducted among more than 3,000 college students from the United States and
Malaysia (Ellis and Hoskin, 2013). In this study, the r2D:4D measure was self-reported.
Although other studies have used a variety of self-reported 2D:4D measures, such
methods are less reliable than studies that use direct physical measurement by trained
researchers (Caswell and Manning, 2009; H ¨
onekopp, 2011; Manning et al., 2010). In
Ellis and Hoskin’s (2013) study, respondents were asked to estimate on a five-point scale
whether the second finger was much longer, slightly longer, essentially the same, slightly
shorter, or much shorter than the fourth finger on their right hand. Despite this crude
method of measurement used, numerous significant inverse correlations between r2D:4D
and criminality within both males and females from both countries were found, possibly
a result of the study’s large sample size. Nevertheless, the imprecision in how r2D:4D was
measured leaves the findings inconclusive.
THIS STUDY
The present study was undertaken to provide a more definitive test of the possibility
that criminality is associated with the r2D:4D finger ratio. Following ENA theory, the spe-
cific hypothesis to be investigated is that individuals with relatively low second to fourth
digit ratios on the right hand (indicating high prenatal androgen exposure) will exhibit
greater involvement in criminal behavior than those with relatively high ratios. Theoreti-
cally, this pattern should hold for both males and females. As explained in more detail in
the next section, we used a precise physical measure of r2D:4D rather than one based on
self-report.
METHODS
Respondents consisted of 190 (43 percent) males and 255 (57 percent) females. The
sample was a combination of undergraduate college students in criminology classes at a
public university in Texas and other volunteers who were at least 18 years of age who
were recruited by these students. Because there was considerable age dispersion in the
6 HOSKIN & ELLIS
current sample (18 to 76 years of age), age was statistically controlled throughout the
main analyses. The racial and ethnic composition of the entire sample was 53 percent
Hispanic, 40 percent non-Hispanic White, and 7 percent African American.
Participants were assured that the original data collected would be 1) kept in a secure
place, 2) identified by number and not by respondent name, 3) examined by research
personnel only, and 4) destroyed as soon as the information was coded and entered into
an electronic file.
MEASURES OF CRIMINALITY
The questionnaire was four pages in length and covered a wide variety of topics, only
a few of which were analyzed for the present study. Aside from basic demographic in-
formation, the variables herein used from the questionnaire involved an inventory of 14
self-reported offenses. Respondents were asked to report how many times during their
lives (if any) they had committed each of these offenses:
1. Serious assault or beating (needing medical treatment)
2. Minor assault or beatings (not needing medical treatment)
3. Sexual assault (including attempted sexual assault, molesting, rape)
4. Domestic or courtship violence
5. Reckless driving (e.g., road bullying)
6. Illegal car racing
7. Serious theft or robbery (including motor vehicle theft)
8. Minor theft or robbery (including shoplifting, purse snatching, and pick pocket-
ing)
9. Serious damage to property (major vandalism and arson)
10. Minor property damage (minor vandalism such as breaking windows in houses
or cars)
11. Bribery, fraud, or other finance-related offenses
12. Distribution of illegal drugs (excluding alcohol)
13. Use of illegal drugs (excluding alcohol)
14. Illegal gambling
The number of times respondents stated they had committed each offense was recorded
unless it exceeded 10; in which case, 10 was recorded. Also, the total types of crimes
committed—a measure of criminal versatility—was computed by assigning each case a
score of 1 for each category of crime with at least one reported offense (and assigning a 0
for no offenses), and then summing across the 14 crime categories.
MEASURING R2D:4D
To measure each respondent’s r2D:4D ratio, we used a desktop printer and scanner
with a 600 ×600 dpi scanner resolution. The respondent’s right hand was placed palm
down on the glass and was held flat to make certain that fingers were equally extended and
to ensure that the scanner captured a clear picture of the entire right hand. GNU Image
Manipulation Program (GIMP 2.8) software was used to measure digitally the length of
the fingers. For both the second and the fourth digits, mouse-controlled calipers were
placed on the basal crease of the finger and extended to the tip of the finger so that the
lengths were measured in number of pixels rounded to one decimal place. The length of
FETAL TESTOSTERONE AND CRIMINALITY 7
the second digit divided by the length of the fourth digit was recorded as the r2D:4D ratio.
Similar methodology has been used by numerous other researchers (e.g., Allaway et al.,
2009; Coolican and Peters, 2003; Fink, Manning, and Neave, 2004; Rahman and Wilson,
2003; Roney and Maestripieri, 2004; van Anders, Vernon, and Wilbur, 2006).
If the two fingers are approximately of equal length (r2D:4D ࣈ1) or if the index finger
is longer (r2D:4D >1), then low fetal testosterone (and high fetal estrogen) is indicated.
If the ring finger is noticeably longer than the index finger (r2D:4D <1), then the testos-
terone levels were high (and estrogen was low) prenatally. For reliability purposes, the
r2D:4D ratio was measured on two different occasions, 3 months apart, by a person who
was unaware of which scan belonged to which questionnaire. Intraobserver reliability
was .89, a coefficient comparable with other studies and indicative of strong reliability
(Allaway et al., 2009).
ADDITIONAL VARIABLES MEASURED
As a closing item on the questionnaire, respondents were asked the following question:
“How carefully did you read and try to honestly answer the questions in this question-
naire?” Response choices ranged from 0 (“not at all honest”) to 10 (“extremely honest”).
The typical respondent reported a high level of honesty and care when answering ques-
tions. On average, the females in the sample stated they had been slightly more honest
(mean =9.37) than males (mean =9.14), although the difference fell slightly short of
being statistically significant (t=–1.84, p=.07).
Other variables that were obtained from the questionnaires were those of sex, age,
and race/ethnicity. The proportionality of race/ethnicity in the sample was as follows:
7 percent Black, 53 percent Hispanic, and 40 percent non-Hispanic White. The gender,
age, race/ethnicity, and honesty measures were included in the multivariate regression
analysis.
DATA ANALYSES
Preliminary analysis revealed that all crime measures except for total crime types were
highly skewed and overdispersed (see table 1). The Spearman rho is helpful with skewed
measures because it reflects similarity of rank ordering, and thus, it was used rather than
the Pearson correlation. The statistical technique adopted for the multivariate analyses
was negative binomial regression. This type of regression is appropriate for models that
use count data for dependent variables that have many incidences of zero offending and
a great deal of dispersion among those who did report committing one or more offenses
(Hilbe, 2011; Kremelberg, 2011). The measure of r2D:4D was standardized to a mean of
zero and standard deviation of 1 for the negative binomial analysis to provide a more
interpretable metric for the results.
RESULTS
Table 1 displays the range of responses, means, and standard deviations for each vari-
able separately by sex. In addition, the sex difference between means for each variable
was assessed for statistical significance with a ttest. The most frequently reported offenses
for both sexes were illegal car racing, reckless driving, and illegal drug use.
8 HOSKIN & ELLIS
Table 1. Descriptive Statistics—Continuous Variables (n=255 Females,
190 Males; N=445)
Males Females
Variable Range Mean (SD) Mean (SD) tScore Significance
Serious assault 0–10 .20 (.93) .15 (.94) .66 .508
Minor assault 0–10 .86 (2.13) .35 (1.46) 2.84 .005
Sexual assault 0–10 .05 (.23) .02 (.12) 2.29 .022
Domestic assault 0–10 .12 (.57) .10 (.67) .36 .716
Serious theft 0–10 .07 (.23) .05 (.23) .58 .565
Minor theft 0–10 .77 (1.91) .51 (1.70) 1.54 .125
Financial crime 0–10 .13 (.80) .05 (.22) 1.35 .180
Serious property damage 0–10 .25 (1.19) .06 (.31) 2.21 .028
Minor property damage 0–10 .77 (2.06) .30 (1.25) 2.80 .006
Illegal car racing 0–10 3.24 (4.32) .69 (2.17) 7.66 .000
Reckless driving 0–10 1.87 (3.24) .69 (2.02) 4.43 .000
Drug dealing 0–10 .88 (2.47) (1.77) (1.77) 2.19 .029
Illegal drug use 0–10 2.74 (4.00) 1.73 (3.48) 2.80 .005
Illegal gambling 0–10 1.42 (3.14) .18 (1.16) 5.21 .000
Total number of offense categories 0–14 2.76 (2.68) 1.42 (2.09) 5.87 .000
r2D:4D ratio .85–1.04 .94 (.03) .96 (.03) –8.25 .000
Age 18–76 31.88 (12.73) 31.73 (12.66) –.12 .903
Honesty measure 1–10 9.14 (1.36) 9.37 (1.20) –1.84 .067
ABBREVIATION:SD=standard deviation.
As expected, the means for males were higher than for females for all 14 categories of
crime as well as for the diversity of crimes committed. Nevertheless, the differences were
not statistically significant for the following five specific crime categories: serious assault,
domestic assault, serious theft, minor theft, and financial crime. Although the absence
of a sex difference for minor theft and perhaps financial crime was not surprising, it was
unexpected for violent crimes like serious and sexual assault.
All categories of crime show striking variation; standard deviations are greater than
their corresponding means in all cases except for the total number of offense categories
for the male sample. Many participants have never committed any one of the 14 types of
crime, whereas a small minority has engaged in each of the crime types numerous times.
Regarding the r2D:4D digit ratio, the mean is significantly higher for females than
for males. In fact, the tscore for the sex difference in r2D:4D (t=–8.25) was higher
than the tscore for any of the sex differences in offending rates. Such a pattern con-
forms to findings from numerous prior studies (e.g., H ¨
onekopp and Watson, 2010; McIn-
tyre, 2006; van Anders, Vernon, and Wilbur, 2006), presumably reflecting the fact that
nearly all females are exposed to lower levels of prenatal testosterone than nearly all
males.
Correlations between finger ratio and self-reported offending are displayed in table 2.
As theoretically predicted, all crime types are significantly and negatively associated with
r2D:4D for the total sample. When the analyses were performed separately according
to sex, all correlations remain negative, although two coefficients for the male sample—
sexual assault and drug dealing—were no longer significant. All correlations for the fe-
male sample were statistically significant.
The magnitude of the relationships for most of the crime categories is in the moderate
range. However, regarding the total crime category (as a measure of offending diver-
sity or versatility) at the bottom of table 2, substantial inverse correlations were found
FETAL TESTOSTERONE AND CRIMINALITY 9
Table 2. Spearman Rank Correlations Between r2D:4D and
Self-Reported Offending for 14 Offense Types for Males,
Females, and the Total Sample
Total Sample Males Females
Offense Types (N=445) (n=190) (n=255)
Serious assault −.255∗∗ −.189∗∗ −.298∗∗
Minor assault −.284∗∗ −.175∗−.299∗∗
Sexual assault −.198∗∗ −.129 −.288∗∗
Domestic assault −.164∗∗ −.179∗−.140∗
Serious theft −.278∗∗ −.245∗∗ −.310∗∗
Minor theft −.258∗∗ −.271∗∗ −.217∗∗
Serious property damage −.217∗∗ −.207∗−.195∗∗
Minor property damage −.242∗∗ −.221∗−.198∗∗
Illegal car racing −.264∗∗ −.250∗∗ −.123∗
Reckless driving −.309∗∗ −.285∗∗ −.222∗∗
Bribery and fraud (financial crime) −.239∗∗ −.190∗∗ −.283∗∗
Drug dealing −.180∗∗ −.062 −.225∗∗
Illegal drug use −.244∗∗ −.178∗−.200∗∗
Illegal gambling −.298∗∗ −.284∗∗ −.188∗∗
Total types of crime −.439∗∗ −.420∗∗ −.446∗∗
∗p<.05; ∗∗p<.01 (two-tailed test).
(i.e., –.420 for males and –.446 for females). This observation indicates that as exposure
to prenatal androgens increases, an individual’s likelihood of committing a wide diversity
of offenses increases considerably.
A summary of our negative binomial regression modeling is shown in table 3, with the
more detailed coefficients for each control variable for all 14 types of offenses presented in
the online supporting information.1Regarding basic interpretation, negative coefficients
indicate that higher prenatal androgen exposure is associated with higher probabilities of
offending (as theoretically predicted). Negative binomial regression coefficients should
be interpreted as the expected log count change in the dependent variable that is asso-
ciated with a one-unit change in the independent variable (Hilbe, 2011). As mentioned
previously, the raw digit ratio scores were standardized to facilitate the interpretation of
the negative binomial regression results.
Table 3 shows the negative binomial regression coefficient for each of the 14 offenses
investigated for 1) the total sample, 2) males only, and 3) females only. When males and
females are analyzed together (the total sample), table 3 demonstrates that the associa-
tions between r2D:4D and offending are statistically significant for all 14 crime categories.
In all cases, the direction is negative (low r2D:4D ratios are associated with higher offend-
ing rates).
When the sample is divided according to sex, the pattern remains, although some coef-
ficients become statistically nonsignificant. For the male sample, all digit ratio coefficients
are negative, but only eight of them reached statistical significance. Those that were not
significant were serious, minor, and sexual assault and all three types of victimless crime
(illicit drug use, drug dealing, and illegal gambling). Among females, r2D:4D is negatively
1. Additional supporting information can be found in the listing for this article in the Wiley Online
Library at http://onlinelibrary.wiley.com/doi/10.1111/crim.2015.53.issue-1/issuetoc.
10 HOSKIN & ELLIS
Table 3. Negative Binomial Regression Model Using r2D:4D Digit
Ratio, Sex, Age, Race/Ethnicity, and Honesty as
Predictors—Only r2D4D Coefficients Are Shown
Total Sample Male Sample Female Sample
(N=445) (n=190) (n=255)
Dependent Variable b(SE) b(SE) b(SE)
Assault
Serious −1.52∗∗∗ (.32) −.78 (.63) −3.37∗∗∗ (.88)
Minor −1.31∗∗∗ (.32) −.40 (.50) −1.27∗∗ (.47)
Domestic −.73∗(.30) −.75∗(.44) −.50 (.43)
Sexual −2.56∗∗∗ (.61) −1.65 (1.16) −3.10∗∗∗ (.79)
Theft
Serious −2.19∗∗∗ (.41) −1.69∗∗ (.56) −2.91∗∗∗ (.77)
Minor −1.17∗∗∗ (.28) −1.11∗∗ (.35) −1.40∗∗ (.48)
Fraud −2.68∗∗∗ (.58) −2.88∗∗ (.89) −2.97∗∗∗ (.81)
Property Damage
Serious −.47∗∗ (.18) −1.81∗(.77) −1.92∗∗ (.66)
Minor −.13∗∗ (.05) −1.13∗(.45) −1.26∗∗ (.45)
Illegal Car Racing −.03∗(.02) −.70∗(.30) −.81 (.65)
Reckless Driving −1.28∗∗∗ (.26) −1.30∗∗∗ (.33) −1.21∗∗ (.42)
Drug Dealing −.58∗(.23) −.25 (.48) −2.17∗∗ (.65)
Illegal Drug Use −.87∗∗∗ (.23) −.22 (.28) −1.52∗∗∗ (.43)
Illegal Gambling −1.09∗∗ (.37) −.60 (.34) −4.43∗(2.11)
ABBREVIATION:SE=standard error.
∗p<.05; ∗∗p<.01; ∗∗∗ p<.001 (two-tailed test).
and significantly associated with 12 of the 14 crime categories; the only exceptions were
domestic assault and illegal racing.
A more detailed negative binomial regression analysis for the offending diversity mea-
sure is presented in table 4. This table clearly shows that having low r2D:4D finger ratios
are associated with elevated involvement in a diversity of crimes. The associations were
strong for the sample as a whole (NBR =–.98) as well as for males (NBR =–.69) and
especially for females (NBR =–1.42) when analyzed separately. Following Paternoster et
al. (1998), we conducted a significance test of the difference between the coefficients for
males and females. The effect for the female sample is significantly larger (z=3.17). Over-
all, having a more male-typical digit ratio (i.e., low r2D:4D) is associated with a greater
diversity of crimes. This finding is true of the whole sample and for each sex, with the
effect being somewhat stronger for women than for men.
Turning to the control variables other than sex (i.e., age, race/ethnicity, and self-
reported honesty), table 4 shows that they are far less important than the r2D:4D variable
(regardless of whether analysis is performed separately by sex). Specifically, years of age
had no significant bearing on the r2D:4D–offending diversity relationships. In the case of
race/ethnicity, the only significant impact was for Hispanic females to report relatively low
involvement in a diversity of offenses compared with White females. Finally, the honesty
item had one significant effect on the r2D:4D–offending diversity relationships. Specifi-
cally, males who professed being the most honest admitted to having committed a greater
diversity of offenses than males who gave low ratings to their honesty in answering the
questionnaire.
FETAL TESTOSTERONE AND CRIMINALITY 11
Table 4. Negative Binomial Regression Model Using r2D:4D Digit
Ratio, Sex, Age, Race/Ethnicity, and Honesty as Predictors of
the Total Categories of Crime Committed
Total Sample Males Females
(N=445) (n=190) (n=255)
Predictors b(SE) b(SE) b(SE)
r2D:4D ratio −.98∗∗∗ (.11) −.69∗∗∗ (.13) −1.42∗∗∗ (.19)
Male .39∗∗∗ (.11) – –
Age .04 (.03) .03 (.03) .08 (.05)
Age-squared/10 −.01 (.00) .00 (.00) −.01∗(.00)
Black .04 (.21) −.11 (.30) −.18 (.32)
Hispanic −.04 (.12) .22 (.14) −.44∗(.20)
Honesty .02 (.04) .11∗(.06) −.08 (.07)
αa.60 .39 0.92
Intercept −.07 (.63) −.78∗∗∗ (.80) −.46 (1.02)
Model χ2126.63∗∗∗ 37.19∗∗∗ 63.87∗∗∗
d.f. 7 6 6
N466 200 266
NOTE: Standard errors are presented in parentheses.
ABBREVIATION: d.f. =degrees of freedom.
aThese values reflect the residual variance in true crime counts, which is dispersion beyond that expected from
a simple Poisson process.
∗p<.05; ∗∗∗p<.001 (two-tailed test).
Again, the overall conclusion to draw from table 4 is that r2D:4D is by far the most
important factor used in the current study to account for variations in offending diversity.
Age, race/ethnicity, and honesty had little-to-no bearing on offending diversity, and the
effects of sex were modest for the sample as a whole (which was of course totally elimi-
nated when males and females were analyzed separately). Finally, it is interesting to note
that the link between r2D:4D and offending diversity was considerably higher for females
(NBR =–1.42) than for males (NBR =–.69; for gender difference in coefficients, z=3.17,
p<.01), suggesting that androgens could be even more pertinent to female variations in
criminality than to male variation.
Because the ages of respondents in our sample varied a great deal and respondents
were asked about lifetime offending, there was some concern that our findings might be
partially a result of participants’ age. To assess this possibility, we not only controlled
for age but also reanalyzed the data only among the respondents who were between 18
and 30 years of age. The pattern of results was very similar to that for the full sample.
In particular, the link between digit ratios and the measure of crime versatility remained
negative and significant for the total younger subsample (NBR =–.87 [.14], p<.001, n=
265), as well as for males (NBR =–.59 [.20], p<.01, n=110) and females (–1.10 [.22],
p<.001, n=155) in the younger subsample.
DISCUSSION
ENA theory rests heavily on lines of evidence that evolution has shaped the functioning
of the human brain so that males can at least partially accommodate female preferences
12 HOSKIN & ELLIS
for mates who are loyal and reliable provisioners of resources (Geary, 2010; Miller, 2011).
These female preferences in turn are thought to have evolved because females with such
preferences have a reproductive advantage over females who use other criteria for choos-
ing mates (Buss and Schmitt, 1993; Ryan, 1998).
According to ENA theory, as a result of females choosing mates who are competent
at resource provisioning, sex hormones such as testosterone have evolved tendencies to
promote competitive and status-striving behavior among males. However, because com-
petitive, status-striving behavior often can victimize others, thereby provoking retaliation
from victims (or their relatives), learning to strive for resources and status in ways that
minimize retaliation is important. As civilizations have evolved, the urges that victims
have to avoid being victimized has given rise to written criminal statutes, which in turn
has brought about the criminal justice system as an enforcer of those statutes (Conley,
1991).
To explain most forms of criminality (including delinquency), ENA theory asserts that
the motivation to compete for resources and status (and for mating opportunities) is
largely hardwired into the brain and that androgens play a central role in guiding this
process (Ellis, 2005). Studies have shown that androgens help to organize several areas
of the brain (Kimura, 1992), particularly in the arousal control centers (Daitzman et al.,
1978) and the neocortex (Wisniewski, 1998; Witelson, 1991).
Androgens also seem to impact at least two important neurotransmitter systems. These
are the dopaminergic system (Hull et al., 1998; Ngun et al., 2011) and the serotoniner-
gic system (van Goozen and Fairchild, 2006). Studies have suggested that the combined
effects of these systems result in greater male tendencies to be competitive, impulsive,
self-directed, and willing to take risks (Christiansen, 1998; Davies, 2014; Smith, Chein,
and Steinberg, 2013; Trent and Davies, 2012). Sex differences in these competitive, im-
pulsive, self-directed, and risk-taking tendencies often are noticeable soon after birth but
normally become accentuated after the onset of puberty when androgen levels surge, par-
ticularly among males (Collaer and Hines, 1995; Hull et al., 1998; Manning et al., 2014).
Another important factor in sexually differentiating the brain—always to varying de-
grees rather than in an either/or fashion—involves the prevalence of androgen receptors
(Ngun et al., 2011). These receptors, located in neurons as well as cells more generally,
are programmed by genetic codes called “CAG repeats” located on the X chromosome
(Allen et al., 1992; Zitzmann and Nieschlag, 2003). To the extent that androgen receptors
are present, many male-typical traits are promoted, including ones of a neurological (and
thereby behavioral) nature (Azurmendi et al., 2006; Kimura, 1992; La Spada et al., 1991;
Sato et al., 2004).
As noted, ENA theory asserts that the motivation for resource acquisition and mating
opportunities are largely unlearned. However, the ability to recognize the most effec-
tive techniques for making these acquisitions interacts heavily with learning ability and
the ability to plan (figure 1). Consequently, as puberty commences, individuals with the
most highly androgenized brains usually exhibit “crude” forms of resource competition
and efforts to mate. Within a few years, most of these individuals develop increasingly
“sophisticated” forms of resource procuring and mating techniques that are within legal
boundaries. In contrast, males who are slow to develop “sophisticated” forms of compe-
tition for resources and mates are those who exhibit the greatest difficulties learning and
planning (Ellis, 2005, 2011). These latter individuals are most likely to exhibit long-term
criminal careers.
FETAL TESTOSTERONE AND CRIMINALITY 13
MAIN RESEARCH FINDINGS
As noted earlier, ENA theory’s central theme is that prenatal brain exposure to an-
drogens has evolved so as to promote competitiveness over resources, status, and mat-
ing opportunities in males, with the stipulation that criminality tends to be crude expres-
sions of this competitiveness. Males should therefore be more involved in crime than fe-
males, as considerable research has documented (reviewed by Ellis, Beaver, and Wright,
2009: 11–7). ENA theory’s prediction of universal sex differences in offending contrasts
with the more prevalent view that sex differences in offending result from social learning
(Akers and Jensen, 2006; Burton et al., 1998; LaGrange and Silverman, 1999; Shekarkhar
and Gibson, 2011; Steffensmeier and Allan, 1996).
In addition to explaining sex differences in criminality in terms of sex differences in
brain exposure to androgens, ENA theory leads to the hypothesis that even within each
gender elevated prenatal androgens should positively correlate with criminality. In other
words, if 1) gender differences in criminal behavior are substantially influenced by pre-
natal androgens and 2) variability exists within both sexes regarding prenatal androgen
exposure—as evidence has suggested (Collaer and Hines, 1995)—then prenatal andro-
gens and criminality should remain correlated even within each sex.
We used negative binomial regression to refine the results of our more basic correla-
tional analysis, with r2D:4D, sex, age, race/ethnicity, and an honesty-in-answering-the-
questionnaire item entered into the prediction equations as control variables. Tables 3
and 4 show that the r2D:4D digit ratio was a strong predictor of involvement in criminal
behavior, and in all cases, the relationship was negative (as theoretically predicted). These
findings confirm the hypothesis that low digit ratios (indicating greater prenatal exposure
to androgens) are associated with increased probabilities of offending.
UNIQUENESS OF THIS STUDY
Although most of the methodology used in this study resembled those reported re-
cently by Ellis and Hoskin (2013), there were two notable differences. First and foremost,
rather than relying on self-reported r2D:4D based on a simple 5-point scale, the present
study used a more reliable objective measure of r2D:4D. This measure has become rec-
ognized as the gold standard measure of the relative lengths of the second and fourth
digits with well-established interrater reliability (e.g., Brown et al., 2002; Grimbos et al.,
2010; H ¨
onekopp and Watson, 2010; Kastlunger et al., 2010; Williams, Greenhalgh, and
Manning, 2003). Furthermore, at least one study has shown that the measure is indeed
indicative of varying prenatal testosterone exposure (Lutchmaya et al., 2004).
Second, although both the Ellis and Hoskin (2013) study and the present one indicated
that low r2D:4D ratios (indicating higher prenatal exposure to androgens) were associ-
ated with self-reported offending, most of findings from the present study were substan-
tially stronger. This result can be explained by noting that in this study a more reliable
physical measure of r2D:4D was used.
As noted in the introduction, three studies have investigated the association between
r2D:4D and officially measured criminality of traffic violation. In this regard, Schwerdt-
feger, Heims, and Heer (2010) found an association of –.36 between males having been
convicted of a moving traffic violation and r2D:4D, and Hanoch et al. (2012) reported that
for male offenders versus male nonoffenders more generally, the association with r2D:4D
was –.24. The magnitude of these associations is similar to the within-male findings from
14 HOSKIN & ELLIS
the present study. Nevertheless, a third study of official involvement in crime by Ander-
son (2012) found no significant differences between males who had been convicted and
those who had not been. Obviously, more research is warranted, but the overall pattern of
findings so far is pointing toward confirmation of the hypothesis that prenatal androgens
have an impact on criminal behavior.
LIMITATIONS
The present study has at least five notable limitations. First, the study was based on a
convenience/snowball sample of university students in Texas majoring in criminology and
friends and family members they recruited for participation in the study. Consequently,
the sample is disproportionately young, middle class, southern, and Hispanic, which limits
the generalizability of our findings and underscores the need for replication.
Second, for some offenses in which sex differences were expected (e.g., serious assault,
domestic assault, and serious theft), none were identified. We believe that this was most
likely because of the failure of our sample to include sufficient numbers of individuals who
were involved in serious criminality. An additional consideration is that only 43 percent
of the sample were males, which further diminished the chances of seriously criminal
individuals having been included in our sample.
Third, the r2D:4D digit ratio has been shown to be a proxy, albeit not very precise,
for prenatal androgen exposure (Hell and P¨
aßler, 2011; Lippa, 2003). Also, one should
bear in mind that factors other than androgens affect digit growth (McIntyre et al., 2005;
Voracek, Manning, and Dressler, 2007). Furthermore, the availability of androgens in the
body’s extremities when a growth spurt occurs in the fourth digit cannot be considered
anything but roughly associated with androgen levels inside the nervous system during crit-
ical periods of fetal development (McIntyre, 2006; Talarovicov´
a, Krskov´
a, and Blazekov´
a,
2009). Also, it is worth noting that self-reported offending provides only a rough gage of
people’s varying tendencies to violate criminal statutes (Moffitt and Silva, 1988; Thorn-
berry and Krohn, 2000). As a result of these limitations in estimating brain exposure to
androgens (as the independent variable) and offending behavior (as the dependent vari-
able), one can assume that the present findings only provide minimal estimates of the
extent to which prenatal brain exposure to androgens affects later involvement in crime.
Fourth, although numerous statistically significant r2D:4D–criminality relationships
were herein documented, and the magnitude of most of these statistical relationships were
modest to moderate, using the r2D:4D ratio for individual-level preoffending diagnostics
or for parole screenings would be inadvisable. We consider the present study’s findings as
mainly being of interest because they provide a clear test of the theoretical deduction that
prenatal androgens are influencing brain functioning patterns in ways that affect people’s
tendencies to commit crime.
Fifth, many hypotheses can be derived from ENA theory, only one of which was tested
by the present study. As illustrated in figure 1, our study focused on determining whether
a positive correlation exists between perinatal androgens (shaded cell to the left) and
criminality (shaded cell to the right). So central is this hypothesis to ENA theory that
failing to confirm it would certainly have put the entire theory into question. Nonetheless,
our having found that prenatal androgens are associated with criminality does not mean
that the rest of the theory is true. Prenatal androgens could be affecting criminal behavior
through pathways that differ substantially from those depicted in figure 1.
FETAL TESTOSTERONE AND CRIMINALITY 15
CONCLUSION
At the heart to ENA theory is the assertion that prenatal brain exposure to testosterone
and other androgens contributes to criminality. Evidence bearing on this hypothesis
involved correlating a biomarker for prenatal androgens known as r2D:4D with self-
reported offending. Overall, the hypothesis that low r2D:4D ratios would be positively
correlated with offending was supported even when the data were examined for males
and females separately, thereby eliminating the possibility that sex differences in both
r2D:4D and offending were confounding the correlations. Nevertheless, further repli-
cations are in order. Also, because ENA theory is complex with numerous component
variables, many other aspects of ENA theory are in need of empirical scrutiny.
REFERENCES
Akers, Ronald L., and Gary F. Jensen. 2006. The empirical status of social learning theory
of crime and deviance: The past, present, and future. In Taking Stock: The Status of
Criminological Theory, eds. Francis Cullen, John Wright, and Kristie Blevins. New
Brunswick, NJ: Transaction.
Allaway, Heather C., Terri G. Bloski, Roger A. Pierson, and Marla E. Lujan. 2009. Digit
ratios (2D:4D) determined by computer-assisted analysis are more reliable than those
using physical measurements, photocopies, and printed scans. American Journal of
Human Biology 21:365–70.
Allen, R. Cutler, Huda Zoghbi, Auda Moseley, Howard Rosenblatt, and John W. Bel-
mont. 1992. Methylation of HpaII and HhaI sites near the polymorphic CAG repeat
in the human androgen-receptor gene correlates with X chromosome inactivation.
American Journal of Human Genetics 51:1229–39.
Anderson, Taja. 2012. Comparing risk taking and digit ratio (2D:4D) in offenders and
non-offenders. Plymouth Student Scientist 5:105–20.
Azurmendi, Aitziber, Francisco Braza, Ainhoa Garc´
ıa, Paloma Braza, Jose M. Mu ˜
noz,
and Jose R. S´
anchez-Mart´
ın. 2006. Aggression, dominance, and affiliation: Their rela-
tionships with androgen levels and intelligence in 5-year-old children. Hormones and
Behavior 50:132–40.
Brown, Windy M., Melissa Hines, Briony A. Fane, and S. Marc Breedlove. 2002. Mas-
culinized finger length patterns in human males and females with congenital adrenal
hyperplasia. Hormones and Behavior 42:380–6.
Burton, Velmer S., Francis T. Cullen, T. David Evans, Leanne F. Alarid, and R. Gregory
Dunaway. 1998. Gender, self-control, and crime. Journal of Research in Crime and
Delinquency 35:123–47.
Buss, David M. 2006. The Murderer Next Door: Why the Mind Is Designed to Kill.
London, U.K.: Penguin Books.
Buss, David M., and David P. Schmitt. 1993. Sexual strategies theory: An evolutionary
perspective on human mating. Psychological Review 100:204–32.
Caswell, Noreen, and John T. Manning. 2009. A comparison of finger 2D:4D by self-
report direct measurement and experimenter measurement from photocopy: Method-
ological issues. Archives of Sexual Behavior 38:143–8.
16 HOSKIN & ELLIS
Christiansen, Kerrin. 1998. Behavioural correlates of testosterone. In Testosterone: Ac-
tion, Deficiency, Substitution, eds. Everhart Nieschlag and Herman M. Behre. New
York: Springer.
Collaer, Marcia L., and Melissa Hines. 1995. Human behavioral sex differences: A role
for gonadal hormones during early development? Psychological Bulletin 118:55–107.
Conley, Carolyn A. 1991. The Unwritten Law: Criminal Justice in Victorian Kent.New
York: Oxford University Press.
Coolican, J., and Michael Peters. 2003. Sexual dimorphism in the 2D/4D ratio and its
relation to mental rotation performance. Evolution and Human Behavior 24:179–83.
Cooper, Jonathan A., Anthony Walsh, and Lee Ellis. 2010. Is criminology moving toward
a paradigm shift? Evidence from a survey of the American Society of Criminology.
Journal of Criminal Justice Education 21:332–47.
Daitzman, Reid, Marvin Zuckerman, Paul Sammelwitz, and Venkalaseshu Ganjam. 1978.
Sensation seeking and gonadal hormones. Journal of Biosocial Science 10:401–8.
Daly, Martin, and Margo Wilson. 1988. Homicide. New Brunswick, NJ: Transaction.
Davies, William. 2014. Sex differences in attention deficit hyperactivity disorder: Candi-
date genetic and endocrine mechanisms. Frontiers in Neuroendocrinology 35:331–46.
DeLisi, Matt. 2011. The limbic system and crime. In The Ashgate Research Companion to
Biosocial Theories of Crime, eds. Kevin M. Beaver and Anthony Walsh. Burlington,
VT: Ashgate.
Ellis, Lee. 2003. Genes, criminality, and evolutionary neuroandrogenic theory. In Bioso-
cial Criminology: Challenging Environmentalism’s Supremacy, eds. Anthony Walsh
and Lee Ellis. New York: Nova Science.
Ellis, Lee. 2005. Theoretically explaining biological correlates of criminal behavior. Eu-
ropean Journal of Criminology 2:287–315.
Ellis, Lee. 2011. Evolutionary neuroandrogenic theory and universal gender differences
in cognition and behavior. Sex Roles 64:707–22.
Ellis, Lee, Kevin M. Beaver, and John Wright. 2009. Handbook of Crime Correlates.Am-
sterdam, the Netherlands: Elsevier/Academic Press.
Ellis, Lee, and Anthony W. Hoskin. 2013. Criminality and the 2D:4D ratio: Testing the
prenatal androgen hypothesis. International Journal of Offender Therapy and Com-
parative Criminology. E-pub ahead of print. doi: 10.1177/0306624×1350381.
Fink, Bernard, John T. Manning, and Nick Neave. 2004. Second to fourth digit ratio and
the “Big Five” personality factors. Personality and Individual Differences 37:495–503.
Geary, D. C. 2010. Male, Female: The Evolution of Human Sex Differences, 2nd ed. Wash-
ington, DC: American Psychological Association.
Ghiglieri, Michael P. 2000. The Dark Side of Man: Tracing the Origins of Male Violence.
New York: Basic Books.
Grimbos, Teresa, Khytam Dawood, Robert P. Burriss, Kenneth J. Zucker, and David A.
Puts. 2010. Sexual orientation and the second to fourth finger length ratio: A meta-
analysis in men and women. Behavioral Neuroscience 124:278–87.
Hanoch, Yaniv, Michaela Gummerum, and Jonathan Rolison. 2012. Second-to-fourth
digit ratio and impulsivity: A comparison between offenders and nonoffenders. PLOS
ONE 7: e47140.
Hell, Benedikt, and Katja P¨
aßler. 2011. Are occupational interests hormonally influ-
enced? The 2D:4D-interest nexus. Personality and Individual Differences 51:376–80.
FETAL TESTOSTERONE AND CRIMINALITY 17
Hilbe, Joseph M. 2011. Negative Binomial Regression, 2nd ed. New York: Cambridge
University Press.
H¨
onekopp, Johannes. 2011. Relationships between digit ratio 2D:4D and self-reported
aggression and risk taking in an online study. Personality and Individual Differences
51:77–80.
H¨
onekopp, Johannes, Luise Bartholdt, Lothar Beier, and Andreas Liebert. 2007. Second
to fourth digit length ratio (2D:4D) and adult sex hormone levels: New data and a
meta-analytic review. Psychoneuroendocrinology 32:313–21.
H¨
onekopp, Johannes, and Sara Watson. 2010. Meta-analysis of digit ratio 2D:4D shows
greater sex difference in the right hand. American Journal of Human Biology 22:619–
30.
Hull, Elaine M., Daniel S. Lorrain, Jainfang Du, Leslie Matuszewich, Daniel Bitran, J.
Ken Nishita, and Laura L. Scaletta. 1998. Organizational and activational effects of
dopamine on male sexual behavior. In Males, Females and Behavior: Toward Biolog-
ical Understanding, eds. Lee Ellis and Linda Ebertz. Westport, CT: Praeger.
Jones, Owen D. 1996. Evolutionary analysis in law: An introduction and application to
child abuse. North Carolina Law Review 75:1117–245.
Kastlunger, Barbara, Stefan G. Dressler, Erich Kirchler, Luigi Mittone, and Martin Vo-
racek. 2010. Sex differences in tax compliance: Differentiating between demographic
sex, gender-role orientation, and prenatal masculinization (2D:4D). Journal of Eco-
nomic Psychology 31:542–52.
Kimura, Doreen. 1992. Sex differences in the brain. Scientific American 267:118–25.
Kremelberg, David. 2011. Practical Statistics. Thousand Oaks, CA: Sage.
LaGrange, Teresa C., and Robert A. Silverman. 1999. Low self-control and opportunity:
Testing the general theory of crime as an explanation for gender differences in delin-
quency. Criminology 37:41–72.
La Spada, Albert R., Elizabeth M. Wilson, Dennis Lubahn, A. E. Harding, and Kenneth
H. Fischbeck. 1991. Androgen receptor gene mutations in X-linked spinal and bulbar
muscular atrophy. Nature 352:77–9.
Lightcap, Joy L., Jeffrey A. Kurland, and Robert L. Burgess. 1982. Child abuse: A test of
some predictions from evolutionary theory. Ethology and Sociobiology 3:61–7.
Lippa, Richard A. 2003. Are 2D:4D finger-length ratios related to sexual orientation? Yes
for men, no for women. Journal of Personality and Social Psychology 85:179–88.
Lutchmaya, Svetlana, Simon Baron-Cohen, Peter Raggatt, Rebecca Knickmeyer, and
John T. Manning. 2004. 2nd to 4th digit ratios, fetal testosterone and estradiol. Early
Human Development 77:23–8.
Manning, John T. 2008. The Finger Book. London: Faber and Faber.
Manning, John T., Simon Baron-Cohen, Sally Wheelwright, and Bernhart Fink. 2010.
Is digit ratio (2D:4D) related to systemizing and empathizing? Evidence from direct
finger measurements reported in the BBC internet survey. Personality and Individual
Differences 48:767–71.
Manning, John, Liam Kilduff, Christian Cook, Blair Crewther, and Bernhard Fink. 2014.
Digit ratio (2D:4D): A biomarker for prenatal sex steroids and adult sex steroids in
challenge situations. Frontiers in Endocrinology 5:9–13.
Manning, John T., Diane Scutt, J. Wilson, and D. Iwan Lewis-Jones. 1998. The ratio of 2nd
to 4th digit length: A predictor of sperm numbers and concentrations of testosterone,
luteinizing hormone and oestrogen. Human Reproduction 13:3000–4.
18 HOSKIN & ELLIS
McIntyre, Matthew H. 2006. The use of digit ratios as markers for perinatal androgen
action. Reproductive Biology and Endocrinology 4:1–9.
McIntyre, Matthew H., Peter T. Ellison, Daniel E. Lieberman, and Bradford Towne.
2005. The development of sex differences in digital formula from infancy in the Fels
Longitudinal Study. Proceedings of the Royal Society B: Biological Sciences 272:1473–
9.
McKibbin, William F., Todd K. Shackelford, Aaron T. Goetz, and Valerie G. Starratt.
2008. Why do men rape? An evolutionary psychological perspective. Review of Gen-
eral Psychology 12:86–97.
Miller, Geoffrey. 2011. The Mating Mind: How Sexual Choice Shaped the Evolution of
Human Nature. New York: Random House.
Moffitt, Terrie E., and Phil A. Silva. 1988. Self-reported delinquency: Results from an
instrument for New Zealand. Australian and New Zealand Journal of Criminology
21:227–40.
Mori, Chisota, and Kohei Shiota. 1994. Sexing of human embryos and fetuses by fluores-
cent in situ hybridization (FISH) to paraffin-embedded tissues with sex chromosome-
specific DNA probes. American Journal of Medical Genetics 50:180–6.
Ngun, Tuck C., Negar Ghahramani, Francisco S´
anchez, Sven Bocklandt, and Eric Vilain.
2011. The genetics of sex differences in brain and behavior. Frontiers in Neuroen-
docrinology 32:227–46.
Paternoster, Raymond, Robert Brame, Paul Mazerolle, and Alex Piquero. 1998. Us-
ing the correct statistical test for the equality of regression coefficients. Criminology
36:859–66.
Rahman, Qazi, and Glenn D. Wilson. 2003. Sexual orientation and the 2nd to 4th fin-
ger length ratio: Evidence for organizing effects of sex hormones or developmental
instability? Psychoneuroendocrinology 28:288–303.
Raine, Adrian. 1997. The Psychopathology of Crime: Criminal Behavior as Clinical Dis-
order. New York: Academic Press.
Roney, James R., and Dario Maestripieri. 2004. Relative digit lengths predict men’s
behavior and attractiveness during social interactions with women. Human Nature
15:271–82.
Ryan, Michel J. 1998. Sexual selection, receiver biases, and the evolution of sex differ-
ences. Science 281:1999–2003.
Sato, Takashi, Takahiro Matsumoto, Hirotaka Kawano, Tomoyuki Watanabe,
Yoshikatsu Uematsu, Keisuke Sekine, Toru Fukuda, Kenichi Aihara, Andree Krust,
Takashi Yamada, Yuko Nakamichi, Yoko Yamamoto, Takashi Nakamura, Kimi-
hiro Yoshimura, Tatsuya Yoshizawa, Daniel Metzger, Pierre Chambon, and Shigeaki
Kato. 2004. Brain masculinization requires androgen receptor function. Proceedings
of the National Academy of Sciences of the United States of America 101:1673–8.
Schwerdtfeger, Andreas, Robert Heims, and Jonannes Heer. 2010. Digit ratio (2D:4D) is
associated with traffic violations for male frequent car drivers. Accident Analysis and
Prevention 42:269–74.
Shekarkhar, Zahra, and Chris L. Gibson. 2011. Gender, self-control, and offending be-
haviors among Latino youth. Journal of Contemporary Criminal Justice 27:63–80.
Smith, Ashley R., Jason Chein, and Lawrence Steinberg. 2013. Impact of socio-emotional
context, brain development, and pubertal maturation on adolescent risk-taking. Hor-
mones and Behavior 64:323–32.
FETAL TESTOSTERONE AND CRIMINALITY 19
Steffensmeier, Darrell, and Emilie Allan. 1996. Gender and crime: Toward a gendered
theory of female offending. Annual Review of Sociology 22:459–87.
Talarovicov´
a, Alzbeta, Lucia Krskov´
a, and Jana Blazekov´
a. 2009. Testosterone enhance-
ment during pregnancy influences the 2D:4D ratio and open field motor activity of rat
siblings in adulthood. Hormones and Behavior 55:235–9.
Thayer, Bradley A., and Valerie M. Hudson. 2010. Sex and the Shaheed: Insights from
the life sciences on Islamic suicide terrorism. International Security 34:37–62.
Thornberry, Terrence P., and Marvin D. Krohn. 2000. The self-report method for mea-
suring delinquency and crime. Criminal Justice 4:33–83.
Thornhill, Randy, and Craig T. Palmer. 2000. A Natural History of Rape. Cambridge, MA:
MIT Press.
Trent, Simon, and William Davies. 2012. The influence of sex-linked genetic mechanisms
on attention and impulsivity. Biological Psychology 89:1–13.
van Anders, Sari, Philip Vernon, and Christopher Wilbur. 2006. Finger-length ratios show
evidence of prenatal hormone-transfer between opposite-sex twins. Hormones and
Behavior 49:315–9.
van de Beek, Cornelieke, Jos H. H. Thijssen, Peggy T. Cohen-Kettenis, Stephanie H. M.
van Goozen, and Jan K. Buitelaar. 2004. Relationship between sex hormones assessed
in amniotic fluid, and maternal and umbilical cord serum: What is the best source
of information to investigate the effects of fetal hormone exposure? Hormones and
Behavior 50:118–25.
van Goozen, Stephanie H. M., and Graeme Fairchild. 2006. Neuroendocrine and neuro-
transmitter correlates in children with antisocial behavior. Hormones and Behavior
50:647–54.
Voracek, Martin, John T. Manning, and Stefan G. Dressler. 2007. Repeatability and inter-
observer error of digit ratio (2D:4D) measurements made by experts. American Jour-
nal of Human Biology 19:142–6.
Walsh, Anthony, and Kevin M. Beaver. 2009. Biosocial Criminology: New Directions in
Theory and Research. New York: Routledge.
Williams, Justin H. G., K. D. Greenhalgh, and John T. Manning. 2003. Second to fourth
finger ratio and possible precursors of developmental psychopathology in preschool
children. Early Human Development 72:57–65.
Wisniewski, Amy B. 1998. Sexually-dimorphic patterns of cortical asymmetry, and the
role for sex steroid hormones in determining cortical patterns of lateralization. Psy-
choneuroendocrinology 23:519–47.
Witelson, Sandra. F. 1991. Neural sexual mosaicism: Sexual differentiation of the hu-
man temporo-parietal region for functional asymmetry. Psychoneuroendocrinology
16:131–53.
Wright, John Paul. 2010. Biosocial Criminology: Oxford Bibliographies Online Research
Guide. New York: Oxford University Press.
Wright, John Paul, and Francis T. Cullen. 2012. The future of biosocial criminology
beyond scholars’ professional ideology. Journal of Contemporary Criminal Justice
28:237–53.
Yang, Yaling, and Adriane Raine. 2009. Prefrontal structural and functional brain imag-
ing findings in antisocial, violent, and psychopathic individuals: A meta-analysis. Psy-
chiatry Research: Neuroimaging 174:81–8.
20 HOSKIN & ELLIS
Zheng, Zhengui, and Martin J. Cohn. 2011. Developmental basis of sexually dimorphic
digit ratios. Proceedings of the National Academy of Science 108:16289–94.
Zitzmann, Michael, and Eberhard Nieschlag. 2003. The CAG repeat polymorphism
within the androgen receptor gene and maleness. International Journal of Andrology
26:76–83.
Anthony W. Hoskin is an associate professor of criminology and criminal justice and
graduate program head at the University of Texas of the Permian Basin. He is interested
in studying offenders and the criminal justice system from a biosocial perspective.
Lee Ellis recently retired after teaching social science for more than 30 years at Minot
State University in Minot, North Dakota, and being a visiting researcher at the University
of Malaya in Kuala Lumpur, Malaysia. His main topics of professional interest involve the
study of criminal behavior, social stratification, and sex differences.
