Digit ratio (2D:4D), dominance, reproductive success, asymmetry, and sociosexuality in the BBC Internet Study

Article (PDF Available)inAmerican Journal of Human Biology 20(4):451-61 · July 2008with436 Reads
DOI: 10.1002/ajhb.20767 · Source: PubMed
Abstract
Digit ratio (2D:4D) may be a correlate of prenatal sex steroids, and has been linked to traits, which are influenced by fetal testosterone and estrogen. Here we consider such links in a large Internet study of sex differences (the BBC Internet Study) in which finger lengths were self-measured. Consistent with lab-based findings the 2D:4D in this study shows sexual dimorphism, ethnic differences and higher dimorphism of right 2D:4D than left, thereby indicating that 2D:4D does measure real between-participant variation. High error in self-measurement of fingers reduces effect sizes. However, the large sample size gives assurance that significant effects are likely to be real. We controlled for ethnicity and sexual orientation by considering White heterosexuals only (153,429 participants). Sexual dimorphism was confirmed in 2D:4D and for the difference of right-left 2D:4D. After Bonferroni correction we found highly significant relationships with low effect sizes as follows. In males and females there were negative associations between 2D:4D and dominance. In males there were negative associations between 2D:4D and family size and factors associated with reproductive success. For females these associations were positive. For asymmetry we found U-shaped relationships with 2D:4D in both males and females. We found no relationship between 2D:4D and promiscuity (sociosexuality). In total, we considered 48 relationships and found 29 to be significant. We compare our findings with a similar study reported by Putz et al. (2004), which found only 2 out of 57 correlations to be significant and discuss possible reasons for the discrepancies between the studies.
Original Research Article
Digit Ratio (2D:4D), Dominance, Reproductive Success, Asymmetry, and
Sociosexuality in the BBC Internet Study
JOHN T. MANNING
1
AND BERNHARD FINK
2
*
1
Department of Psychology, University of Central Lancashire, Preston, United Kingdom
2
Department of Sociobiology/Anthropology, Institute of Zoology and Anthropology, University of Go¨ttingen, Go¨ttingen, Germany
ABSTRACT Digit ratio (2D:4D) may be a correlate of prenatal sex steroids, and has been linked to traits, which are
influenced by fetal testosterone and estrogen. Here we consider such links in a large Internet study of sex differences
(the BBC Internet Study) in which finger lengths were self-measured. Consistent with lab-based findings the 2D:4D in
this study shows sexual dimorphism, ethnic differences and higher dimorphism of right 2D:4D than left, thereby indi-
cating that 2D:4D does measure real between-participant variation. High error in self-measurement of fingers reduces
effect sizes. However, the large sample size gives assurance that significant effects are likely to be real. We controlled
for ethnicity and sexual orientation by considering White heterosexuals only (153,429 participants). Sexual dimorphism
was confirmed in 2D:4D and for the difference of right–left 2D:4D. After Bonferroni correction we found highly signifi-
cant relationships with low effect sizes as follows. In males and females there were negative associations between
2D:4D and dominance. In males there were negative associations between 2D:4D and family size and factors associated
with reproductive success. For females these associations were positive. For asymmetry we found U-shaped relation-
ships with 2D:4D in both males and females. We found no relationship between 2D:4D and promiscuity (sociosexuality).
In total, we considered 48 relationships and found 29 to be significant. We compare our findings with a similar study
reported by Putz et al. (2004), which found only 2 out of 57 correlations to be significant and discuss possible reasons for
the discrepancies between the studies. Am. J. Hum. Biol. 20:451–461, 2008.
' 2008 Wiley-Liss, Inc.
The relative length of the second (index) and fourth
(ring) fingers (digit ratio or 2D:4D) is sexually dimorphic
such that males—on average—have lower mean values of
2D:4D than females. That is, in males the ring finger is
usually longer than the index finger, whereas in females
the length difference is smaller (Manning et al., 1998;
Manning, 2002). This dimorphism has been known for
more than 100 years (Baker, 1888; reviewed in Peters
et al., 2002), but over this period little work concerning its
developmental origins has been undertaken. By consider-
ing the work of Phelps (1952) and Garn et al. (1975) on the
ontogeny of finger length determination Manning et al.
(1998) suggested that the sex difference in digit ratio orig-
inates in utero under the influence of fetal testosterone
(FT) and fetal estrogen (FE). It was this seminal sugges-
tion that stimulated much recent work on 2D:4D and has
led to the publication of some 180 papers since 1998.
There have been a number of studies designed to test
the hypothesis that the sexual dimorphism in 2D:4D is
formed under the influence of FT and FE. Supportive data
are now available but conclusive proof is not yet forthcom-
ing. Briefly, the evidence for a link between 2D:4D and
prenatal sex steroids includes manipulation experiments
of bird 2D:4D by adding sex steroids to eggs (Romano
et al., 2005; Saino et al., 2007), data from measurements
of the early sex dimorphism of human 2D:4D in utero
(Lutchmaya et al., 2004; Malas et al., 2006) and in chil-
dren (McIntyre et al., 2005, 2006), studies of 2D:4D in
children with congenital adrenal hyperplasia (Brown
et al., 2002), comparisons between 2D:4D in females with
a male twin and 2D:4D of females with a female twin (van
Anders et al., 2006), reports of the effects on 2D:4D of poly-
cystic ovary syndrome (Catrall et al., 2005) and of mater-
nal smoking (Rizwan et al., 2007), and a study of 2D:4D
and sexual dimorphism in the brain (Kallai et al., 2005).
FT and FE may have wide-ranging organizing effects on
aspects of sex-dependent behavior (Collaer and Hines,
1995; Reinisch et al., 1991). Consistent with such an
effect, Manning (2002) set out preliminary evidence that
2D:4D was a significant correlate of many human traits
that show sex differences in the areas of fertility, behavior,
and health. However, contrary to Manning’s (2002) argu-
ments, Putz et al. (2004) reported little association
between 2D:4D and traits such as status, mating success,
components of reproductive success, asymmetry, spatial
ability, sociosexuality, and sexual orientation. Indeed
these authors found that only two out of 57 correlations
tested were significant and in the predicted direction, and
three were significant but counter to the predicted direc-
tion. On the basis of these results Putz et al. (2004) con-
cluded that the usefulness of 2D:4D as a predictor of sex-
hormone-mediated traits was limited.
The ‘‘British Broadcasting Corporation (BBC) Internet
Study’’ [hereafter termed ‘‘Study’’] provides a unique op-
portunity to re-evaluate links between 2D:4D and many of
the traits considered by Manning (2002) and Putz et al.
(2004) in a very large sample. There have been reports
from the Study on relationships between 2D:4D and sex-
ual orientation (Manning et al., 2007), and in two tests of
spatial judgment (Collaer et al., 2007; Peters et al., 2007).
The conclusions were not consistent with those of Putz
*Correspondence to: Bernhard Fink, Department of Sociobiology/An-
thropology, Institu te of Zoology and Anthropology, University of Go
¨
ttingen,
Berliner Strasse 28, D-37073 Go
¨
ttingen, Germany.
E-mail: bernhard.fink@ieee.org
Received 23 August 2007; Revision received 27 November 2007;
Accepted 13 January 2008
DOI 10.1002/ajhb.20767
Published online 23 April 2008 in W iley InterScience (www.interscience.
wiley.com).
AMERICAN JOURNAL OF HUMAN BIOLOGY 20:451–461 (2008)
V
V
C
2008 Wiley-Liss, Inc.
et al. (2004). One possible reason for the discrepancies
may lie in the different modes of finger measurement. The
Study uses direct measurement of finger length while
Putz et al. (2004) reported indirect measurements of fin-
ger length from photocopies. There is evidence that the
indirect method distorts 2D:4D downwards (Manning
et al., 2005) and may have led to misleading results in
many studies concerned with 2D:4D and sexual orienta-
tion (Manning et al., 2007). With this possibility in mind
we consider relationships in the BBC data between 2D:4D
and other traits (sexual dimorphism, dominance, repro-
ductive success, asymmetry, sociosexuality), which were
reported in Putz et al. (2004).
The Study was designed to investigate sex differences
in a large web-based sample of people all over the world. A
detailed description of the Study together with a full list
of questions is given in Reimers (2007). Web-based studies
give us the opportunity to collect data from very large
samples, but there are fewer controls on accuracy than
are possible in lab-based studies. Nevertheless, web- and
lab-based studies give remarkably similar results in the
majority of cases (reviewed in Krantz and Dalal, 2000).
There are two ways of checking data reliability within
web-based studies i.e. checking internal consistency and
investigating consistency with other studies. Evidence for
consistency with other studies and for internal consis-
tency within the Study has been discussed by Reimers
(2007). Cross-checks were done on such items as gender
and contraceptive pill use, age and occupation, age and
education, and sexual orientation and attraction to males
and females. Overall, the percentage of participants giv-
ing inconsistent answers tended to be from <1 to 5%. For
example, of the participants who said they were male
0.9% claimed they were taking the contraceptive pill. Sim-
ilarly, of the participants who said they were under 16
years 2.1% claimed they had a University education. The
consistency with previous studies was checked for a num-
ber of traits. For example as expected male advantage was
shown in mental rotation performance (Peters et al.,
2007) and judgment of line orientation performance (Col-
laer et al., 2007). Also consistent with other studies, MRT
scores were higher in women taking the contraceptive pill
and in lesbians compared to heterosexual women, and
homosexual men scored lower for MRT than heterosexual
men (Peters et al., 2007).
As part of the Study, participants were asked to mea-
sure their second and fourth fingers on the right a nd
left hands and report this via an internet web-page.
With regard to 2D:4D, the consistency of the data in
the Study with previous studies of 2D:4D has been con-
sidered by Manning et al. (2007). 2D:4D is sexually
dimorphic such that males have mean 2D:4D, which is
lower than mean female 2D:4D (Manning et al., 1998;
Manning, 2002), the sex difference is universal in
humans but 2D:4D also shows e thnic variation with,
for example, lower mean 2D:4D in Black populations
compared to Caucasians (Manning, 2002; Manning
et al., 2000), and the sex difference in 2D:4D is greater
in right hands compared to left hands (Manning et al.,
1998). Consistent with previous studies it was found
that 2D:4D was lower in males than females, the sex
dif ference was present in all ethnic groups, there were
significant ethnic differences with Black participants
showing lower mean 2D:4D than Caucasians, and the
sexual dimorphism in 2D:4D was stronger in the right
hand compared to the left hand (Manning et al., 2007).
Furthermore, a considerat ion of mean right and left
2D:4D for White participants across 13 countries
showed in all but one comparison males had lower
2D:4D than females and the sex differences in 11 com-
parisons were significant. In addition, the sexual
dimorphism in the r ight hand was significantly greater
than that of the left hand (Manning et al., 2007). It is
therefore p lausible to assume that self-measured 2D:4D
reported in the B BC Study doe s indeed capture suffi-
cient of the between-participant variation in 2D:4D to
show established sex, ethnic and hand differences in
2D:4D. In addition Caswell and Manning (in press)
have reported that self-measured finger lengths are
similar to experimenter-measured finger lengths. How-
ever, the sex-dependent effect sizes of self-reported
2D:4D are lower than effect sizes of 2D:4D obtained
from lab-based measurement. Weak effect sizes may be
countered by large sample sizes. The BBC Study
recruited a total of some 250,000 participants and even
after adjustment for ethnic and other differences in the
Study there remains a large sample (Manning et al.,
2007). The Caswell and Manning study showed a tend-
ency for a few values of 2D:4D (about 1%) to be very
high or very low due to major errors or deliberately
mislea ding responses. The frequency and magnitude of
these outliers were similar in the BBC Study and the
Caswell and Manning report. In the latter, removal of
self-measured outliers of 2D:4D increased the concord-
ance of 2D:4D between self- and lab-based measure-
ments. The procedure of removing outliers of 2D:4D in
the BBC Study was followed by Manning et al. (2007),
Manning and Quinton (2007), Peters et al. (2007) and
Collaer et al. (2007).
The purpose of this present study was two-fold. Firstly,
we considered the BBC Study in relation to existing evi-
dence for links between 2D:4D and sexual dimorphism,
dominance, reproductive success, secondary factors asso-
ciated with reproductive success, asymmetry and socio-
sexuality. Secondly, we compared the 2D:4D associations
found in the BBC Study with the results reported by Putz
et al. (2004). Thus, the fundamental aim of this study was
to evaluate 2D:4D as a potential biomarker for sex-de-
pendent behavior by clearing the way for promising future
research directions in this area.
METHODS
The BBC Study
The questions in the Study are discussed and listed in
Reimers (2007), and the Study itself can be seen at URL:
http://www.bbc.co.uk/science/humanbody/sex. The Study
took on average about 40 min per participant to complete.
There were six sections containing in total about 200
questions, four cognitive tasks, and other tasks such as
judging emotions.
Participants
Descriptive statistics of the sample are given by Reim-
ers (2007). Briefly, there were a total of 462,859 partici-
pants who completed the initial demographic information
and 255,116 participants who completed the Study. Of the
latter, 52.7% were males and 47.3% females. With regard
to ethnicity, 84.1% described themselves as White, 6.3% as
452
J.T. MANNING AND B. FINK
American Journal of Human Biology
Asian or Asian British, 2.2% as Chinese, 1.2% as Middle
or Near Eastern, 0.8% as Black or Black British, 0.8% as
Black other, and the remainder (2.2%) as of mixed origin.
The sample was very diverse in comparison to many lab-
based studies of university students. Participants
reported some 25 occupations and were from more than
100 countries. However, the largest proportion of respond-
ents lived in the UK (46.8%) and the USA (27.7%). With
regard to age, the mean was 29.39 6 11.91 years, with a
very small percentage ( <0.3%) saying they were 10 years
or younger or 70 years or older. The questions on sexual
orientation were restricted to participants who were 18
years. Of those who responded, 90% stated they were het-
erosexual (straight), 4.2% homosexual (gay or lesbian)
and 5.5% bisexual.
Finger length measurements
Participants were asked to measure their index finger
and ring finger on the palm of the right and left hand (see
Appendix for instructions). For each finger there was a
drop-down menu from 10 to 100 mm inclusive. It was felt
that rounding to the nearest millimeter would capture
sufficient real between-individual variation in the data
given the large sample size.
Restrictions on the data
We restricted our analyses to participants who were
18 years of age, White, heterosexual, and with 2D:4D in
the right and left hands, which was within the range of
0.80 to 1.20. The reasons for these restrictions were as
follows:
Measurements of finger lengths in the Study were
taken directly from the fingers. There is evidence that
direct measurements of the fingers are more accurate
than measurements taken from photocopies of fingers,
and 2D:4D calculated from the former tends to be higher
than 2D:4D from the latter (Manning et al., 2005). How-
ever, finger lengths that are self-measured occasionally
include measurement errors that result in extreme values
of 2D:4D. Caswell and Manning (in press) found that self-
measured values of 2D:4D which were outside the range
of 0.80 to 1.20 showed evidence of mistakes which
made them very different to experimenter-measured
2D:4D. Manning et al. (2007) considered 2D:4D from
direct measurements (Manning et al., 1998) and found
that the range of 0.80 to 1.20 included almost all lab-
based values of 2D:4D. The adoption of this range of
2D:4D from the Study had the effect of removing only 1%
of the participants from each tail of the distribution.
Therefore, in order to eliminate extreme values we fol-
lowed protocols of previously published work on 2D:4D in
the Study and considered values in the range of 0.80 to
1.20 (Collaer et al., 2007; Manning and Quinton, 2007;
Manning et al., 2007; Peters et al., 2007).
There were known effects of ethnicity on 2D:4D in the
Study (Manning et al., 2007). Whites, the majority group,
had higher mean 2D:4D than Blacks and Chinese. Also in
the Study there were significant effects of sexual orienta-
tion on 2D:4D in males (Manning et al., 2007). Therefore,
we restricted our analyses to White heterosexual partici-
pants.
With regard to age, the questions concerning sexual
behavior in the Study were presented to participants
who were 18 years of age. Therefor e, we excluded par-
ticipants from the statistical analyses who were <18
years old.
Correction for adult fetal sex steroids
Jasienska et al. (2006) have reported negative correla-
tions of finger asymmetry with estradiol levels in women.
Estradiol may be linked to potential fertility and other
traits examined here. Therefore, we adjusted our results
for finger asymmetry, and for the purposes of comparison
have provided associations, which are independent of
asymmetry.
Correction for multiple tests
Studies of associations between 2D:4D and target traits
usually report associations with right and left 2D:4D, and
occasionally with the mean of both. The examination of
multiple correlations results in an increased likelihood of
false positive relationships. Many reports of 2D:4D, there-
fore, apply Bonferroni corrections when testing relation-
ships (e.g. Manning, 2002). However, Bonferroni correc-
tion assumes that associations are independent of one
another, but right and left 2D:4D are quite strongly corre-
lated. In the present study we considered associations
between right and left 2D:4D and target traits and
decided to apply Bonferroni correction. However, we
emphasize this is a conservative procedure and probably
removes from consideration a number of relationships
that show real effects. In addition, mean 2D:4D is highly
correlated with right and left 2D:4D and gives little more
information. Studies with large numbers of non-signifi-
cant associations for right and left 2D:4D run the risk of
inflating the number of non-significant relationships by
considering mean 2D:4D (e.g. Putz et al., 2004). Therefore,
we did not consider mean 2D:4D in the present analyses.
Predicted results
We base our predictions wherever possible from studies
that have measured finger length directly rather than
indirectly from photocopies. With regard to sexual dimor-
phism, we expected 2D:4D and right–left 2D:4D (Dr-l)
would be lower in males than females (direct finger mea-
surement, Manning, 2002; Manning et al., 1998). For the
target traits we predicted: (i) 2D:4D would be negatively
correlated with dominance measures because dominance
scores are likely to be higher in males than females and
this may be due to differences in PT; (ii) reproductive suc-
cess in males would be negatively related to 2D:4D and in
females positively related to 2D:4D as reported by Man-
ning et al. (2000) and Manning (2002) in samples using
direct finger measurements. It should be noted here that
these effects were present in societies where family size
restriction was common (e.g. the UK) and not common
(e.g. Zulu townships in South Africa). This suggests that
factors other than fertility (e.g. desire for children) may be
related to 2D:4D; (iii) following from (ii) in the absence of
evidence to the contrary factors related to reproductive
success would be negatively related to 2D:4D in males and
positively related to 2D:4D in females; (iv) finger asymme-
try would be related to 2D:4D in a curvilinear U-shaped
way as reported by Manning et al. (2006) in a number of
samples using direct finger measurement; (v) measures of
sociosexuality would be negatively related to 2D:4D
453
DIGIT RATIO (2D:4D) IN THE BBC INTERNET STUDY
American Journal of Human Biology
because they are usually higher in males than females
and this may result from higher male PT. Details of the
target traits in the Study are as follows (see Appendix for
the actual items).
Sexual dimorphism
Sexual dimorphism in 2D:4D in the Study has been con-
sidered by Manning et al. (2007). The first question of the
Study concerned sex and all participants were required to
answer this before they were allowed to continue.
Dominance
Dominance in the Study was assessed from ten items
taken from the International Personality Pool (IPIP;
http:/ipip.ori.org). Participants responded to each item
on a seven point scale from disagree 5 1 to agree 5 7.
Therefore, possible scores could range from 10 to 70.
Participants scoring less than 10 must have missed
answering one or more items and therefore had received
a zero for that item(s). Scores that indicated such miss-
ing items were excluded and a range of scores from 10
to 70 considered.
Reproductive success
The Study included an item ‘‘How many children do you
have?’’ Participant’s response was recorded in a drop-
down menu with scores ranging from 0 to 10 in single
steps and 101.
Factors related to reproductive success
The Study included items that may indirectly relate to
reproductive success (see Appendix) including preferred
number of children, age at first child, preferred age at first
child and how much the participant wants to have or care
for children, frequency of sex, strength of the sex drive
and how easily participants are sexually excited.
Asymmetry
Right–left asymmetry was available for the index finger
and ring finger. We obtained signed asymmetry (R–L) for
each finger and removed the signs to obtain absolute
asymmetry (|R–L|). To control for the effect of absolute
finger length on asymmetry we calculated relative
asymmetry for each finger: Relative asymmetry ¼
jR Lj
ðR þ LÞ 3 0:5 To remove very extreme cases from
the data we restricted the range of 2D and 4D asymmetry
from perfect symmetry (0) to 10% asymmetry (0.1).
Following the procedure of Putz et al. (2004) we calculated
the relative asymmetries for the index and ring finger
then summed these and divided by 2 to obtain the compos-
ite asymmetry (Asy
2,4
).
The sociosexual inventory
The sociosexual inventory (SOI) measures promiscuity,
i.e. the willingness to engage in uncommitted sexual rela-
tions (Simpson and Gangestad, 1991). Low scores indicate
a restricted or monogamous mating strategy and high
scores an unrestricted promiscuous mating strategy. The
Study contained seven items from which an SOI score
could be calculated. The first three items reflect overt soci-
osexual variation. They include questions on number of
sexual partners in the last year, the number of these that
were on one occasion only, and the number the participant
estimates having sex with in the next 5 years. Item 4 con-
cerns covert sociosexual behavior. It asks whether the par-
ticipant fantasizes about having sex with someone other
than their partner. Items 5–7 measure sociosexual atti-
tudes. They include ‘‘Sex without love is ok,’’ ‘‘I can imag-
ine myself being comfortable and enjoying casual sex with
different partners,’’ and ‘‘I would have to be closely
attached (emotionally and psychologically) to someone
before I’d feel comfortable and fully enjoy having sex with
him/her.’’
To obtain a measure of promiscuity per participant the
above scores are summed. However, the score from each
item is weighted. The total SOI composite score was calcu-
lated in the following way (see Schmitt, 2005 and Simpson
and Gangestad, 1991): SOI 5 (5 3 item 1) 1 (5 3 item 2)
1 (1 3 item 3 [with a cap of 30]) 1 (4 3 item 4[reverse
scored]) 1 (2 3 mean of items 5, 6, and 7 [reverse scored]).
RESULTS
Sample size
After controlling for ethnicity, sexual orientation and
removing outliers our total sample was 153,429 partici-
pants (84,000 males). When considering each target traits
this figure was reduced somewhat as a result of factors
such as non-completion of some items (e.g. see dominance)
or because the item did not apply to all participants (e.g.
see age at first child).
Sexual dimorphism
Here, we are concerned with sexual dimorphism in right
and left 2D:4D and in right–left 2D:4D (Dr-l) in White het-
erosexual participants. We found significantly lower
2D:4D and Dr-l in 84,000 males compared with 69,429
females (Table 1). The largest effect size of the sexual
dimorphism was for right 2D:4D, then left 2D:4D and the
weakest effect size was for Dr-l. The significant sexual
dimorphism of Dr-l in the Study has not been reported
previously.
Dominance
There was a significant sex difference with medium
effect size in dominance scores, with higher mean scores
(SD) in males compared to females (males 41.92 [10.59], n
5 83,689, females 35.83 [11.67], n 5 69,109, t 5 80.91, P
5 0.0001, d 5 0.54).
The relationships (Pearson correlations) between 2D:4D
and dominance scores are reported in Table 2. There were
significant negative relationships in both males and
females, i.e. participants with low 2D:4D scored higher on
TABLE 1. The effect of sex on right and left 2D:4D and right–left
2D:4D (Dr-l) in male (n 5 84,000) and female (n 5 69,429) White
heterosexual participants
Males Females
dMean SD Mean SD
Right 2D:4D 0.984 0.049 0.994 0.051 20.20*
Left 2D:4D 0.985 0.047 0.993 0.049 20.17*
Dr-l 20.001 0.040 0.002 0.043 20.07*
*P 5 0.0001, significant after Bonferroni correction.
454
J.T. MANNING AND B. FINK
American Journal of Human Biology
dominance than participants with high 2D:4D (see Fig. 1).
The associations were significant for right and left hands,
but right 2D:4D was more strongly related to dominance
than left 2D:4D. The relationships remained significant
after Bonferroni correction. Effect sizes (r
2
from 0.0001 to
0.0007) and P values (0.003 to 0.0001) were both low. Cor-
rection for finger asymmetry reduced the P value for one
out of four associations.
Reproductive success
The reported number of children ranged from 0 to 101,
with 0.2% of males and 0.1% of females reporting seven or
more children. As expected the mean number of children
did not significantly differ between males and females
(males 0.69 [1.17], n 5 83,681, females 0.68 [1.13], n 5
69,173, t 5 1.73, P 5 0.24).
The relationships between 2D:4D and reproductive suc-
cess with the influence of age removed are shown in Ta-
ble 3. There were significant negative associations
between 2D:4D and number of children in males and sig-
nificant positive associations in females (see Fig. 2).
Removing the participants who reported seven or more
children made little difference to these associations. The
relationships remained significant after Bonferroni correc-
tion. Effect sizes (r
2
from 0.00005 to 0.0003) and P values
(0.01 to 0.0001) were both low. Correction for finger asym-
metry reduced the P value for two out of four associations.
Factors related to reproductive success
The preferred number of children ranged from 0 to 101.
As with number of children there were no sex differences
in the number of children preferred (males 2.61 [2.39], n
5 83,072, females 2.59 [2.23], n 5 68,574, t 5 1.93, P 5
0.054).
The range for age at first child and preferred age at first
child was 0 to 99 years. We set a more realistic range of
16–50 years for both items. Females were younger than
males at first child, with medium effect size (males 28.37
[5.44] years, n 5 26,872, females 25.14 [5.27] years, n 5
22,369, t 5 50.60, P 5 0.0001, d 5 0.59), and in their pre-
ferred age at first child, with small to medium effect size
(males 29.67 [4.35] years, n 5 61,535, females 28.05 [4.13]
years, n 5 53,384, t 5 49.37, P 5 0.0001, d 5 0.37).
Preference scores to have and care for children ranged
from 1 to 7. Females reported a stronger preference than
males, with very small effect size (males 4.18 [1.65], n 5
83,107, females 4.21 [1.79], n 5 68.541, t 5 2.99, P 5
0.003, d 5 0.02).
With regard to frequency of sexual intercourse and li-
bido, the scores for frequency of sex ranged from 1 to 6,
and for sex drive and ease of sexual excitement from 1 to
7. Males reported significantly higher frequencies of sex
than females with very small effect size (males 3.29 [1.26],
n 5 83,040, females 3.23 [1.22], n 5 68,120, t 5 2.61, P 5
0.009, d 5 0.05). Males also reported a significantly stron-
ger sex drive than females (males 5.40 [1.49], n 5 81,964,
females 4.61 [1.77], n 5 67,034, t 5 71.33, P 5 0.0001),
and a significantly stronger tendency to get sexually
excited (males 5.46 [1.39], n 5 81,962, females 4.33 [1.71],
n 5 66,863, t 5 104.54, P 5 0.0001) (see Fig. 3). These lat-
ter two items had medium to large effect sizes for the sex
differences (sex drive d 5 0.48; sexual excitement d 5
0.73).
The relationships between 2D:4D and factors associated
with reproductive success are reported in Table 4 for
males and Table 5 for females. In total there were 28 asso-
ciations tested and 18 were significant (9 for males and 9
for females). Earlier reports have suggested that low
2D:4D was related to high reproductive success in males,
and high 2D:4D was associated with high reproductive
success in females (Manning, 2002). We performed Bon-
ferroni corrections across males and females separately.
After correction for multiple tests there were 12 signifi-
cant correlations, six in males and eight in females. Of
these 12 associations controlling for finger asymmetry
reduced the P values for four traits and increased the P
values for two traits. We report only these 12 associations.
In men, low 2D:4D was related to correlates of high
reproductive success. Thus, in comparison to males with
high 2D:4D, males with low 2D:4D reported a stronger
Fig. 1. The relationship between right 2D:4D (with standard error
bars) and dominance scores in 83,689 men (negative association) and
69,109 women (negative association). The dominance scores are
grouped as follows: 1 5 10–19, 2 5 20–29, 3 5 30–39, 4 5 40–49, 5 5
50–59, 6 5 60–70. Sample sizes are shown for each group of domi-
nance scores.
TABLE 2. Relationships between right and left 2D:4D and dominance scores in 83,689 males and 69,109 females
Males Females
rtP rtP
Right 2D:4D 20.015 (20.015) 4.42 (4.15) 0.0001* (0.0001) 20.026 (20.026) 6.86 (6.82) 0.0001* (0.0001)
Left 2D:4D 20.010 (20.012) 2.93 (3.40) 0.003* (0.0007) 20.019 (20.020) 5.06 (5.28) 0.0001* (0.0001)
For comparison, the values in brackets are relationships between 2D:4D and dominance after the removal of the influence of finger asymmetry (81,713 males and
67,049 females).
*Significant after Bonferroni correction.
455
DIGIT RATIO (2D:4D) IN THE BBC INTERNET STUDY
American Journal of Human Biology
preference for children (right and left 2D:4D), a higher
frequency of sex (left 2D:4D), a higher sex drive (left
2D:4D), and a stronger tendency to get sexually excited
(right and left 2D:4D).
In women, high 2D:4D was related to correlates of high
reproductive success. Thus, in comparison with females
with low 2D:4D, females with high 2D:4D reported an ear-
lier age at first child (right and left 2D:4D), an earlier pre-
ferred age at first child (right 2D:4D and left 2D:4D), a
higher sex drive (right and left 2D:4D), and a stronger
tendency to get sexually excited (right and left 2D:4D).
After Bonferroni correction, for both males and females,
effect sizes (r
2
from 0.0001 to 0.002) and P values (0.002 to
0.0001) were both low.
It is of interest to know whether 2D:4D is associated
with number of children independent of factors associated
with reproductive success. For males left 2D:4D is most
strongly linked to family size after the effect of age was
removed. Left 2D:4D was also significantly associated
with preferred number of children, frequency of sex, sex
drive and easily sexually excited. We removed the effect of
age and the latter factors associated with reproductive
success and found left 2D:4D remained significantly nega-
tively associated with number of children (n 5 80,185,
standardized coefficient 520.008, t 5 2.86, P 5 0.004).
Likewise, with female right 2D:4D we removed the effect
of age, preferred age at first child, sex drive and easily sex-
ually excited and found that right 2D:4D remained signifi-
cantly positively associated with numbers of children (n 5
51,628, standardized coefficient 5 0.006, t 5 2.03, P 5
0.04).
Asymmetry
There was a significant sex difference in Asy
2,4
of small
effect size (d 5 0.12), with females showing higher asym-
metry than males (males Asy
2,4
5 0.018 [0.016], n 5
82,015, females Asy
2,4
5 0.020 [0.017], n 5 67,361, t 5
16.34, P 5 0.0001). The relationships between 2D:4D and
Asy
2,4
are shown in Table 6. There were positive and sig-
nificant linear relationships, such that both males and
females with high 2D:4D had high Asy
2,4
. However, there
were stronger U-shaped curvilinear relationships between
2D:4D and Asy
2,4
in males (right 2D:4D, y 5 0.48–0.935 3
x 1 0.473 3 x
2
; left 2D:4D, y 5 0.36–0.7 3 x 1 0.358 3 x
2
)
and females (right 2D:4D, y 5 0.586–1.138 3 x 1 0.57 3
x
2
; left 2D:4D, y 5 0.485–0.931 3 x 1 0.464 3 x
2
). That is,
high values of asymmetry were shown by male and female
participants with low and high values of 2D:4D. The cur-
vilinear relationships with Asy
2,4
were stronger for right
hand 2D:4D than left in both males and females. Effect
sizes (r
2
from 0.09 to 0.15) and P values (all 0.0001) were
both low.
SOI
There was a significant sex difference in SOI scores
with males scoring higher than females (males 36.32
Fig. 3. The relationships between mean left 2D:4D (with standard
error bars) and scores for the item ‘‘It doesn’t take much to get me sex-
ually excited’’ (1 5 disagree to 7 5 agree) in 81,962 men (negative
association) and 66,863 women (positive association). Sample size is
shown for each score for sexual excitement.
TABLE 3. Relationships between right and left 2D:4D and reproductive success in 83,681 males and 69,173 females
Males Females
rtP rtP
Right 2D:4D 20.007 (20.008) 2.54 (2.77) 0.01* (0.006) 0.018 (0.017) 5.80 (5.47) 0.0001* (0.0001)
Left 2D:4D 20.011 (20.012) 4.01 (4.18) 0.0001* (0.0001) 0.010 (0.010) 3.09 (3.14) 0.002* (0.002)
For comparison, the values in brackets are relationships between 2D:4D and reproductive success after the removal of the influence of finger asymmetry (81,703
males and 67,117 females).
*Significant after Bonferroni correction.
Fig. 2. The relationships between mean right 2D:4D (with stand-
ard error bars) and numbers of children in 83,506 men (negative asso-
ciation) and 69,092 women (positive association). Mean 2D:4Ds for
participants with 7–11 children are not shown as there were fewer
than 100 per group and error bars were very large. Sample size is
shown for each family size.
456
J.T. MANNING AND B. FINK
American Journal of Human Biology
[87.88], n 5 64,954, females 26.59 [77.64], n 5 48,165, t 5
14.54, P 5 0.0001, d 5 0.12).
The relationships between 2D:4D and SOI scores are
shown in Table 7. There was no evidence for an associa-
tion between 2D:4D and a tendency towards promiscuity
in either males or females. After adjustment for finger
asymmetry one association became significant (P 5 0.01).
However, as this relationship was counter to the predicted
direction we think this correlation requires clarification
with further work.
DISCUSSION
We have the following results:
i. Right and left 2D:4D showed significant sexual dimor-
phism such that males had lower mean 2D:4D than
females. In addition Dr-l also showed significant sex-
ual dimorphism with lower values for males compared
with females.
ii. Dominance scores showed sex differences, with higher
scores in males. There were significant negative corre-
lations between right and left 2D:4D and dominance
scores in both men and women.
iii. Reproductive success showed no significant sex differ-
ence in number of children. For males there were sig-
nificant negative associations between 2D:4D and
number of children (right and left 2D:4D), and for
females there were positive correlations between
2D:4D and number of children (right and left 2D:4D).
All associations were independent of age.
iv. Considering correlates of reproductive success. There
were no sex differences in preferred number of chil-
dren. However, females were significantly younger at
first child and in their preferred age at first child than
males, and males reported that they had sex more of-
ten, had higher sex drive and were more easily sexu-
ally excited than females. In general, males showed
negative associations between 2D:4D and correlates of
reproductive success. After correction for multiple
tests there were significant negative relationships for
preferred number of children (right and left 2D:4D),
experience of sex (left 2D:4D), strength of sex drive
(left 2D:4D), and ease of attaining sexual excitement
(right and left 2D:4D). In contrast to males, females
tended to show positive associations between 2D:4D
and correlates of reproductive success. After Bonfer-
roni correction there were significant positive associa-
tions between 2D:4D and age at first child (right and
left 2D:4D), preferred age at first child (right and left
2D:4D), sex drive (left 2D:4D), and ease of attaining
sexual excitement (right and left 2D:4D).
v. With regard to 2D:4D and asymmetry, we found signif-
icant positive correlations, but U-shaped relationships
were stronger than linear associations. After Bonfer-
roni correction both men and women showed signifi-
cant U-shaped correlations between 2D:4D and a com-
posite asymmetry score for the second and fourth fin-
gers (right and left 2D:4D).
vi. We found promiscuity (as measured by the SOI) was
not related to right or left 2D:4D for either men or
women.
TABLE 4. Relationships between right and left 2D:4D and variables relating to reproductive success in males
N
Right 2D:4D Left 2D:4D
rtP rtP
How many children would
you like?
83,072 (81,108) 20.011 (20.012) 3.18 (3.49) 0.002* (0.0005) 20.012 (20.012) 3.39 (3.44) 0.0007* (0.0006)
Age at first child 26.872 (26,321) 0.001 (0.0005) 0.12 (0.08) 0.90 (0.94) 0.004 (0.003) 0.62 (0.43) 0.53 (0.67)
Preferred age at first child 61,535 (60,277) 20.006 (20.008) 1.49 (1.91) 0.14 (0.06) 20.002 (20.002) 0.62 (0.49) 0.54 (0.63)
How much do you want to
care for children?
83,107 (81,149) 20.006 (20.004) 1.61 (1.23) 0.11 (0.22) 20.009 (20.008) 2.65 (2.35) 0.008 (0.02)
Intake/experience of sex 83,040 (81,092) 20.009 (20.007) 2.63 (1.89) 0.009 (0.06) 20.013 (20.009) 3.87 (2.58) 0.0001* (0.01)
Sex drive 81,964 (80,037) 20.009 (20.01) 2.61 (2.76) 0.009 (0.006) 20.012 (20.013) 3.57 (3.54) 0.0004* (0.0004)
Easily sexually excited 81,962 (80,032) 20.014 (20.014) 4.03 (3.82) 0.0001* (0.0001) 20.013 (20.015) 3.85 (4.29) 0.0001* ((0.0001)
For comparison, the values in brackets are relationships between 2D:4D and variables relating to reproductive success after the removal of the influence of finger
asymmetry.
*Significant after Bonferroni correction.
TABLE 5. Relationships between right and left 2D:4D and variables relating to reproductive success in females
N
Right 2D:4D Left 2D:4D
rtP r tP
How many children would
you like
68,574 (66,530) 20.001 (20.0004) 0.38 (0.11) 0.70 (0.92) 20.0001 (20.0001) 0.01 (0.01) 0.99 (0.90)
Age at first child 22,369 (21,702) 20.039 (20.040) 5.90 (5.97) 0.0001* (0.0001) 20.026 (20.027) 3.93 (3.98) 0.0001* (0.0001)
Preferred age at first child 53,384 (51,859) 20.028 (20.03) 6.51 (6.80) 0.0001* (0.0001) 20.022 (20.022) 5.14 (5.12) 0.0001* (0.0001)
How much do you want to
care for children
68,541 (66,494) 0.009 (0.010) 2.41 (2.53) 0.02 (0.01) 0.006 (0.007) 1.54 (1.71) 0.13 (0.09)
Intake/Experience of sex 68,120 (66,102) 20.005 (20.006) 1.35 (1.59) 0.18 (0.11) 20.004 (20.006) 1.14 (1.61) 0.25 (0.11)
Sex drive 67,034 (65,032) 0.010 (0.011) 2.71 (2.68) 0.007* (0.007) 0.011 (0.010) 2.83 (2.65) 0.005* (0.01)
Easily sexually excited 66,863 (64,872) 0.012 (0.014) 2.99 (3.65) 0.003* (0.0003) 0.015 (0.015) 3.88 (3.87) 0.0001* (0.0001)
For comparison, the values in brackets are relationships between 2D:4D and variables relating to reproductive success after the removal of the influence of finger
asymmetry.
*Significant after Bonferroni correction.
457
DIGIT RATIO (2D:4D) IN THE BBC INTERNET STUDY
American Journal of Human Biology
vii. We confirmed that the sexual dimorphism of right
2D:4D was greater than that of left 2D:4D (Manning
et al., 2007). However, there was no clear pattern of
‘‘side effects’’ in relationships with target traits. Out
of 29 significant relationships there were 14 right side
correlations and 15 left side correlations. For three
traits one side was significantly associated with a tar-
get trait and one not (in two the significant side was
left [males], and in one the significant side was right
[females]). However, in all three traits the ‘‘non-signif-
icant’’ side had P values ranging from 0.003 to 0.009,
which became non-significant only after adjustment
for multiple tests.
Large Internet studies often show weak effect sizes (Col-
laer et al., 2007; Krantz and Dalal, 2000; Reimers, 2007),
and our findings are no exception to this trend. Self-mea-
sured finger lengths are prone to substantial error and
our reported effect sizes are likely to be substantially
lower than the ‘‘real’’ effect sizes (Caswell and Manning,
in press). However, the sample sizes of the Study are very
large and give us cause for confidence in establishing
whether relationships are real.
As expected, 2D:4D was sexually dimorphic with lower
2D:4D in males compared with females (Manning, 2002;
Manning et al., 1998). The finding that Dr-l was signifi-
cantly lower in men than women probably settles the
question of whether Dr-l is sexually dimorphic. Manning
(2002) reported a similar significant sex difference in a
White sample, but a non-significant difference in a sample
of mixed ethnicities (Manning, 2002). It is likely that the
latter result was unreliable because ethnicity was not
controlled for in this sample. In total there were 52 rela-
tionships examined between 2D:4D and our target varia-
bles. Of these, 36 were significant after Bonferroni adjust-
ment. The significant associations with 2D:4D were in the
expected direction for dominance (negative; Manning,
2002), reproductive success and correlates of reproductive
success (negative for males and positive for females;
Manning et al., 2000; Manning, 2002), and asymmetry (U-
shaped; Manning et al., 2006).
If 2D:4D is a marker for the effect of prenatal sex ste-
roids on the development of various sex dependent traits
and behavior, as originally suggested by Manning et al.
(1998), our findings in a large sample of males and
females support the hypotheses that high PT and/or low
PE is related to dominance in men and women, reproduc-
tive success and correlates of reproductive success in men.
Low PT and/or high PE is associated with reproductive
success and correlates of reproductive success in women.
High PT and/or high PE is related to high asymmetry in
men and women. We emphasize that in these instances
the findings from the Study are in accord with lab-based
reports and with theory. As such they cannot be explained
as artifacts of random measurement error in self-report of
finger lengths.
The Study links between 2D:4D and family size confirm
the findings of Manning et al. (2000). It should be noted
that 2D:4D has previously been found to be related to
reproductive success in populations with both high and
low rates of birth control (Manning, 2002; Manning et al.,
2000). The Study data relate to a predominantly Wester-
nized sample in which birth control would be expected to
be common. This suggests that the 2D:4D links with fertil-
ity may be associated with the desire to have children and
a high sex drive in addition to the efficient functioning of
the reproductive system. Indeed we have found evidence
in the Study that 2D:4D is significantly linked to these
traits. Overall our findings support the suggestion that
genes for high PT and high PE are maintained in the pop-
ulation because they are sexually antagonistic (Manning,
2002; Manning et al., 2000). That is, high PT is advanta-
geous in male fetuses in that it confers reproductive
advantages in men. Genes for high PT are deleterious
TABLE 6. Linear and curvilinear relationships between 2D:4D and composi te asymmetry (Asy
2,4
) of the index and ring fingers in
males (n 5 82,015) and females (n 5 67,361)
Males Females
rt Prt P
Product–moment correlation
Right 2D:4D 0.019 5.42 0.0001* 0.033 8.56 0.0001*
Left 2D:4D 0.034 9.88 0.0001* 0.012 3.00 0.003
rF P rF P
Second-order polynomial analyses
Right 2D:4D 0.12 595.75 0.0001* 0.15 778.15 0.0001*
Left 2D:4D 0.09 366.11 0.0001* 0.12 454.91 0.0001*
*Significant after Bonferroni correction.
TABLE 7. Relationships between right and left and 2D:4D and SOI scores in males (n5 64,954) and females (n5 48,165)
Males Females
rtPrtP
Right 2D:4D 20.002 (20.003) 0.60 (0.72) 0.55 (0.47) 0.007 (0.012) 1.47 (2.55) 0.14 (0.01)
Left 2D:4D 0.0005 (0.005) 0.13 (1.19) 0.90 (0.24) 0.006 (0.005) 1.32 (0.99) 0.19 (0.32)
For comparison, the values in brackets are relationships between 2D:4D and SOI scores after the removal of the influence of finger asymmetry (63,912 males and
46,756 females).
458
J.T. MANNING AND B. FINK
American Journal of Human Biology
when they are present in female fetuses. Similarly, high
PE in female fetuses confers reproductive advantages in
women. Genes for high PE are deleterious when they are
present in male fetuses.
A comparison between the results of Putz et al. (2004)
and the BBC Study findings of Manning et al. (2007),
Peters et al. (2007), Collaer et al. (2007) and the present
report shows many discrepancies. Apart from agreement
that 2D:4D is not related to sociosexuality, there is dis-
cordance in almost all other aspects of the findings. It is
unclear why most of these discrepancies have arisen but
some may be the result of (i) methodological issues con-
cerning 2D:4D, and (ii) methodological issues concerning
the target variables.
Methodological issues concerning 2D:4D
Both studies found sexual dimorphism in 2D:4D such
that males had lower 2D:4D than females. However, Putz
et al. (2004) reported mean 2D:4D (male 2D:4D, right
0.949, left 0.955; female 2D:4D, right 0.974, left 0.979),
which was substantially lower than the mean 2D:4D for
White heterosexual males and females (Table 1) in our
present Study. This suggests that the 2D:4Ds reported by
Putz et al. (2004) are too low and are affected by the dis-
tortional influence of photocopies (Manning et al., 2005;
Manning et al., 2007). A similar finding was reported by
Manning et al. (2007) in a comparison of Study and photo-
copy 2D:4D in reports of sexual orientation. Furthermore,
the strength of the distortional effect of photocopies varies
across the right and left hand (Manning et al., 2005). This
‘‘hand effect’’ may give rise to two problems. Firstly, asym-
metry of R–L finger lengths may be distorted such that
finger asymmetries calculated from photocopies are not
reliable (Manning et al., 2006). Secondly, R–L differences
in 2D:4D (Dr-l) may be affected. Putz et al. (2004) reported
a non-significant sex difference for Dr-l, but a significant
sexual dimorphism was found in the present study.
Adding to the potential problems caused by measuring
2D:4D from photocopies, Putz et al. did not control for the
strong ethnicity effects on 2D:4D (Manning et al., 2000;
Manning et al., 2007). Populations in the US are made up of
many ethnic groups and neglecting to control for this when
measuring 2D:4D can potentially mislead investigators.
Methodological issues concerning the
measurement of target variables
If target variables are not measured accurately real
associations with 2D:4D may be obscured. In the Putz
et al. study, there are concerns regarding the measure-
ment of dominance and mating success. Dominance was
measured not from a multi-item questionnaire as in the
present Study, but from a single item in which partici-
pants were asked to rate their own dominance on a 6-point
scale. With regard to mating success, the Putz et al. sam-
ple measured number of potential conceptions in young
undergraduates. Many of the participants would not be
expected to have children at that point in their lives. In
such a sample it is difficult to predict future reproductive
success with confidence.
There may also have been sample problems in the mea-
surement of sexual orientation. In their Study 2, Putz
et al. considered a sample of 239 participants (119 males)
that contained an undisclosed number of homosexual
males and females. In such a sample, there are likely to be
very few actual homosexual participants. Without knowl-
edge of the relative numbers of heterosexuals and homo-
sexuals it is difficult to evaluate the analysis. Their conclu-
sion, that male homosexuals and female homosexuals have
lower 2D:4D than male or female heterosexuals (Williams
et al., 2000) is not supported by the Study findings. Man-
ning et al. (2007) reported Study findings of higher 2D:4D
for male homosexuals compared to heterosexuals and no
difference in 2D:4D for lesbians and female heterosexuals.
In conclusion, our results from White heterosexuals in
the BBC Study suggest that 2D:4D and Dr-l are sexually
dimorphic with low values (high prenatal T) in males com-
pared to higher values in females. With regard to the tar-
get traits, 2D:4D was significantly negatively associated
with scores for dominance and spatial ability in men and
women, suggesting a link between high PT and dominance
and spatial ability. There were also significant negative
relationships between 2D:4D, family size and factors
related to reproductive success in males, indicating high
PT is implicated in direct male fitness. In women, family
size and factors related to reproductive success tended to
be significantly positively associated with 2D:4D. This lat-
ter finding suggests that low PT and/or high PE are
related to direct fitness in females. Low and high 2D:4D
were linked to high asymmetry in males and females, sug-
gesting that high levels of prenatal sex steroids (PT or PE)
cause developmental instability, which is in accord with
the findings of Manning et al. (2006) for body asymmetry
and Fink et al. (2004) for facial asymmetry. The sample
sizes are large giving us confidence that the associations
are real, but effect sizes in all these relationships are not
clear because self-measurement of finger length is prone
to random error.
We think many stu die s of 2D:4D are methodologically
flawed. The measurement of 2D:4D involves accuracy
in differentiating small differences between the second
and fourth finger lengths of the order of a few milli-
meters. Future studies should calculate 2D:4D from
lab-based direct finger measurements in order to clarify
the effect sizes of the associations we have found in the
Study. Where possible the use of photocopies should be
avoided. We acknowledge that photoc opi es are, in some
instances, the only practical mode of recording finger
lengths. However, where there are no such practical
constraints, e.g. where the participants are readily ac-
cessible as is the case for university students, fingers
should be measured directly. In particular, asymmetry
of fingers must not be obtained from photocopies. Such
asymmetries are unreliable and should be avoided.
Finally, ethnicity differences in 2D:4D are as large as
sex differences. It is essential that ethnicity is con-
trolled for in future 2D:4D studies.
LITERATURE CITED
Baker F. 1888. Anthropological notes on the human hand. Am Anthropol
1:51–76.
Brown WM, Hines M, Fane BA, Breedlove SM. 2002. Masculinized finger
length patterns in human males and females with congenital adrenal
hyperplasia. Horm Behav 42:380–6.
Caswell N, Manning JT. A comparison of finger 2D:4D by self-report direct
measurement and experimenter measurement from photocopy: meth-
odological issues. Arch Sex Behav, in press.
Cattrall FR, Vollenhoven BJ, Weston GC. 2005. Anatomical evidence for
in utero androgen exposure in women with polycystic ovary syndrome.
Fertil Steril 84:1689–1692.
459
DIGIT RATIO (2D:4D) IN THE BBC INTERNET STUDY
American Journal of Human Biology
Collaer ML, Hines M. 1995. Human behavioural sex differences: a role for
gonadal hormones during early development? Psychol Bull 118:55–
107.
Collaer ML, Reimers S, Manning JT. 2007. Visuospatial performance on an
internet line judgement task and potential hormonal markers: sex, sex-
ual orientation, and 2D:4D. Arch Sex Behav 36:177–192.
Fink B, Manning JT, Neave N, Grammer K. 2004. Second to fourth digit
ratio and facial asymmetry. Evol Hum Behav 25:125–132.
Garn SM, Burdi AR, Babler WJ. 1975. Early prenatal attainment of adult
metacarpal–phalangeal rankings and proportions. Am J Phys Anthropol
43:327–332.
Jasienska G, Lipson SF, Ellison PT, Thune I, Ziomkiewicz A. 2006. Sym-
metrical women have higher potential fertility. Evol Hum Behav
27:390–400.
Kallai J, Csatho
´
A, Ko
¨
ve
´
r F, Maka
´
ny T, Nemes J, Horva
´
th K, Kova
´
cs N,
Manning JT, Nadel L, Nagy F. 2005. MRI-assessed volume of left and
right hippocampi in females correlates with the relative length of the
second and fourth fingers (the 2D:4D ratio). Psychiatry Res 140:199–
210.
Krantz JH, Dalal R. 2000. Validity of Web-based psychological research.
In: Birnbaum MH, editor, Psychological experiments on the internet.
San Diego: Academic Press. p 35–60.
Lutchmaya S, Baron-Cohen S, Raggatt P, Knickmeyer R, Manning JT.
2004. Second to 4th digit ratios, fetal testosterone and estradiol. Early
Hum Dev 77:23–28.
Malas MA, Dogan S, Hilal Evcil E, Desdicioglu K. Fetal development of the
hand, digits and digit ratio (2D:4D). Early Hum Dev 82:469–475.
Manning JT. 2002. Digit ratio: a pointer to fertility, behavior, and health.
New Brunswick, NJ: Rutgers University Press.
Manning JT, Barley L, Walton J, Lewis-Jones DI, Trivers RL, Singh D,
Thornhill R, Rohde P, Bereckei T, Henzi P, Soler M, Szwed, A. 2000. The
2nd:4th digit ratio, sexual dimorphism, population differences, and
reproductive success: evidence for sexually antagonistic genes? Evol
Hum Behav 21:163–1 83.
Manning JT, Churchill AJG, Peters M. 2007. The effects of sex, ethnicity
and sexual orientation on self-measured digit ratio (2D:4D). Arch Sex
Behav 36:223–233.
Manning JT, Fink B, Neave N, Caswell N. 2005. Photocopies yield lower
digit ratios (2D:4D) than direct finger measurements. Arch Sex Behav
34:329–333.
Manning JT, Fink B, Neave N, Szwed A. 2006. The 2nd to 4th digit ratio
and asymmetry. Ann Hum Biol 33:480–492.
Manning JT, Quinton S. 2007. Association of digit ratio (2D:4D) with self-
reported attractiveness in men and women: Evidence from the BBC
internet survey. J Indiv Diff 28:73–77.
Manning JT, Scutt D, Wilson J, Lewis-Jones DI. 1998. The ratio of 2nd to
4th digit length: a predictor of sperm numbers and concentrations of tes-
tosterone, luteinizing hormone and oestrogen. Hum Reprod 13:3000–
3004.
McIntyre MH, Cohn BA, Ellison PT. 2006. Sex dimorphism in digital for-
mulae of children. Am J Phys Anthropol 129:143–150.
McIntyre MH, Ellison PT, Lieberman DE, Demerath E, Towne B. 2005.
The development of sex differences in digital formula from infancy in
the Fels Longitudinal Study. Proc Biol Sci 270:2167–2172.
Peters M, Mackenzie K, Bryden P. 2002. Finger length and distal finger
extent patterns in humans. Am J Phys Anthropol 117:209–217.
Peters M, Manning JT, Reimers S. 2007. The effects of sex, sexual orienta-
tion, and digit ratio (2D:4D) on mental rotation performance. Arch Sex
Behav 36:251–260.
Phelps VR. 1952. Relative index finger length as a sex-influenced trait in
man. Am J Hum Genet 4:72–89.
Putz DA, Gaulin SJC, Sporter RJ, McBurney DH. 2004. Sex hormones and
finger length: what does 2D:4D indicate? Evol Hum Behav 25:82–199.
Reimers S. 2007. The BBC Internet Study: general methodology. Arch Sex
Behav 36:147–161.
Reinisch JM, Ziemba-Davis M, Sanders SA. 1991. Hormonal contributions
to sexually dimorphic behavioral development in humans. Psychoneur-
oendocrinology 16:213–278.
Rizwan S, Manning JT, Brabin BJ. 2007. Maternal smoking during preg-
nancy and possible effects of in utero testosterone: evidence from the
2D:4D finger length ratio. Early Hum Dev 83:87–90.
Romano M, Rubolini D, Martinelli R, Bonisoli Alquati A, Saino N. 2005.
Experimental manipulation of yolk testosteron e affects digit length
ratios in the ring-necked pheasant (Phasianus colchicus). Horm Behav
48:342–346.
Saino N, Romano M, Boncoraglio G. 2007. Increased egg estradiol concen-
tration feminizes digit ratios of male pheasants (Phasianus colchicus).
Naturwissenschaften 94:207–212.
Schmitt DP. 2005. Sociosexuality from Argentina to Zimbabwe: a 48-nation
study of sex, culture, and strategies of human mating. Beh Brain Sci
28:247–311.
Simpson JA, Gangestad SW. 1991. Individual differences in sociosexuality:
evidence for convergent and discriminant validity. J Pers Soc Psychol
60:870–883.
van Anders SM, Vernon PA, Wilbur CJ. 2006. Finger-length ratios show
evidence of prenatal hormone-tran sfer between opposite-sex twins.
Horm Behav 49:315–319.
Williams TJ, Pepitone ME, Christensen SE, Cooke BM, Huberman AD,
Breedlove NJ, Breed love TJ, Jordan CL, Breedlove SM. 2000. Finger-
length ratios and sexual orientation. Nature 404:455–456.
APPENDIX
Instructions for measurement of fingers
‘‘For this task you will need a ruler. Hold your right
hand in front of you. Look at where your ring finger joins
the palm of your hand. Find the bottom crease. Go to the
middle of this crease. Put the 0 of your ruler exactly on
the middle of the bottom crease. Make sure the ruler runs
straight up the middle of your finger. Measure to the tip of
your finger (not your nail) in millimeters. Every milli-
meter counts so it is important to do this as accurately as
possible—you are contributing to real scientific research
and it will determine your results.
Participants then responded in a drop-down menu (10
mm, 11 mm ... 100 mm) for the right hand index and ring
finger and the left hand index and ring finger.
Sex
‘‘Before you start, please answer the following: Gender
(Male or Female).’’
Dominance
The Study asked the question ‘‘How do you see yourself’’
and then listed the following items (http:/ipip.ori.org/
newCPIKey.htm#Dominance):
Try to surpass others’ accomplishments.
Try to outdo others.
Am quick to correct others.
Impose my will on others.
Demand explanations from others.
Want to control the conversation.
Am not afraid of providing criticism.
Challenge others’ point of view.
Lay down the law to others.
Put people under pressure.
Participants responded on a seven-point scale from ‘‘dis-
agree’’ to ‘‘agree.’’
Reproductive success
How many children do you have? Drop down menu,
none to 101.
Components of reproductive success
How many children would you like to have (or have
had)? Free format text field.
What age were you when you had your first child? Free
format text field.
What age would you like to have (or have had) your first
child? Free format text field.
How much do you want to have or care for children com-
pared with others of your age or sex? Scale of 1–7, 1 5 not
at all, to 7 5 very.
460
J.T. MANNING AND B. FINK
American Journal of Human Biology
In addition, there were items that related to the frequency
of sexual intercourse, and the strength of the sex drive:
How would you describe your intake/experience of sex?
1 5 never, 2 5 rarely, 3 5 monthly, 4 5 weekly, 5 5 daily,
6 5 more than once a day.
I have a strong sex drive. Scale of 1–7, 1 5 strongly dis-
agree, to 7 5 strongly agree.
It doesn’t take much to get me sexually excited. Scale of
1–7, 1 5 strongly disagree, to 7 5 strongly agree.
SOI
The Study contained seven items from which an SOI
score could be calculated:
How many partners have you had sexual intercourse
with in the past year? Free format text field.
How many of these were only on one occasion? Free for-
mat text field.
With how many partners do you for esee yourself
having sex with in the next 5 years? Free format text
field.
If you have a partner—how often do you fantasize about
having sex with someone other than your partner? Drop
down menu, 1 5 every day, 2 5 few times a week, 3 5
once a week, 4 5 once every three weeks, 5 5 once a
month, 6 5 never.
How much do you agree with the following: sex without
love is ok? Scale of 1–7, 1 5 strongly disagree, to 7 5
strongly agree.
I can imagine myself being comfortable and enjoying
casual sex with different partners. Scale of 1–7, 1 5
strongly disagree, to 7 5 strongly agree.
I would have to be closely attached (emotionally and
psychologically) to someone before I’d feel comfortable and
fully enjoy having sex with him/her. Scale of 1–7, 1 5
strongly disagree, to 7 5 strongly agree.
461
DIGIT RATIO (2D:4D) IN THE BBC INTERNET STUDY
American Journal of Human Biology
    • "Consistent with our specific hypotheses, feminist activists exhibited a significantly more masculinized 2D:4D ratio relative to both Swedish and aggregate comparison groups, a substantially higher level of Directiveness than both the male and female Australian samples, and within-sample correlations between these variables. Consistent with previous research, there were also significant correlations between the hands (Hampson et al., 2008; Manning and Fink, 2008; Butovskaya et al., 2010) and stronger correlations "
    [Show abstract] [Hide abstract] ABSTRACT: The feminist movement purports to improve conditions for women, and yet only a minority of women in modern societies self-identify as feminists. This is known as the feminist paradox. It has been suggested that feminists exhibit both physiological and psychological characteristics associated with heightened masculinization, which may predispose women for heightened competitiveness, sex-atypical behaviors, and belief in the interchangeability of sex roles. If feminist activists, i.e., those that manufacture the public image of feminism, are indeed masculinized relative to women in general, this might explain why the views and preferences of these two groups are at variance with each other. We measured the 2D:4D digit ratios (collected from both hands) and a personality trait known as dominance (measured with the Directiveness scale) in a sample of women attending a feminist conference. The sample exhibited significantly more masculine 2D:4D and higher dominance ratings than comparison samples representative of women in general, and these variables were furthermore positively correlated for both hands. The feminist paradox might thus to some extent be explained by biological differences between women in general and the activist women who formulate the feminist agenda.
    Full-text · Article · Sep 2014
    • "Thus, our data support one of the outlined approaches, showing that women with rather masculine body traits also tend to have a higher number of sexual partners and start earlier with sexual activities, and have thus more unrestricted sexual behavior (Clark, 2004; Mikach and Bailey, 1999). On the other hand, we also supported the other possibility (Manning and Fink, 2008; Rahman et al., 2005), by showing that rather more feminine women have higher sociosexual desires. Sociosexual desires do not need to reflect behaviors (Penke and Asendorpf, 2008), but they can increase, for example, flirting behavior with men who are willing to provide immediate resources or status. "
    [Show abstract] [Hide abstract] ABSTRACT: One of the possible explanations for human within-sex variation in promiscuity stems from conditional strategies dependent on the level of body sex-dimorphism. There is some evidence that masculine men and feminine women are more promiscuous than their sex-atypical counterparts, although mixed results persist. Moreover, another line of evidence shows that more promiscuous women are rather sex-atypical. We tested whether diverse sex-dimorphic body measures (2D:4D, WHR/WSR, handgrip strength, and height and weight) influence sociosexual desires, attitudes, promiscuous behavior, and age of first intercourse in a sex-typical or sex-atypical direction. Participants were 185 young adults, 51 men and 54 women from Brazil, and 40 men and 40 women from the Czech Republic. In men stronger handgrip and more feminine 2D:4D predicted higher sociosexual behaviors, desires, and lower age of the first sexual intercourse. While in women, sociosexual desires were predicted by lower handgrip strength and more feminine 2D:4D. It thus seems that it is rather a mixture of masculine and feminine traits in men, and feminine traits in women that increase their sociosexuality. Masculine traits (height) predicting female promiscuous behavior were specific for only one population. In conclusion, a mosaic combination of sex-typical but also sex-atypical independent body traits can lead to higher promiscuity, particularly in men. Limitations, implications, and future directions for research are considered.
    Full-text · Article · Aug 2014
    • "reproductive success, pre-natal hormonal environment (reflected in 2D:4D value) seems to play a role as well. Variables associated with male fertility such as sperm number (Manning et al., 1998), sex drive, level of sexual excitement (Manning & Fink, 2008), number of sexual partners per individual (Honekopp et al., 2006) and age at first marriage (Sorokowski et al., 2012) were all related to 2D:4D. These results might suggest that digit ratio might be widely associated with traits related to male reproductive strategy. "
    [Show abstract] [Hide abstract] ABSTRACT: Objectives: The 2nd to 4th digit ratio (2D:4D) is thought to reflect exposure to androgens during foetal development. This study examined the relationship between low (more masculine) and high (more feminine) 2D:4D and body size at different stages of the life course, adult testosterone levels and number of children among males. Methods: Five hundred and fifty-eight men from rural Poland at the Mogielica Human Ecology Study Site participated in this study. Life history data and anthropometric measurements were collected. Salivary morning and evening testosterone levels among 110 men from the same population were measured. Results: Low 2D:4D was related to higher birth weight (p = 0.04), higher birth length (p = 0.01), higher body mass during childhood and adolescence (p = 0.01), higher BMI (borderline significance, p = 0.06), higher number of children among fathers (p = 0.04) and higher testosterone levels during adulthood (p = 0.04). Conclusions: This study shows, for the first time in a single population, that digit ratio is related to sub-adult body size at different stages of the life course, adult testosterone levels and number of children. The observed results suggest that digit ratio might be a valuable predictor of male body size and reproductive characteristics.
    Full-text · Article · Apr 2014
Show more