Summarize Psychology Article

Research Report

The Cutest Little Baby Face A Hormonal Link to Sensitivity to Cuteness in Infant Faces R. Sprengelmeyer,1 D.I. Perrett,1 E.C. Fagan,1 R.E. Cornwell,1 J.S. Lobmaier,1 A. Sprengelmeyer,2

H.B.M. Aasheim,1 I.M. Black,1 L.M. Cameron,1 S. Crow,1 N. Milne,1 E.C. Rhodes,1 and A.W. Young3

1 University of St. Andrews,

2 Universität Bielefeld, and

3 University of York

ABSTRACT—We used computer image manipulation to de-

velop a test of perception of subtle gradations in cuteness

between infant faces. We found that young women (19–26

years old) were more sensitive to differences in infant

cuteness than were men (19–26 and 53–60 years old).

Women aged 45 to 51 years performed at the level of

the young women, whereas cuteness sensitivity in women

aged 53 to 60 years was not different from that of

men (19–26 and 53–60 years old). Because average age at

menopause is 51 years in Britain, these findings suggest the

possible involvement of reproductive hormones in cuteness

sensitivity. Therefore, we compared cuteness discrimina-

tion in pre- and postmenopausal women matched for age

and in women taking and not taking oral contraceptives

(progestogen and estrogen). Premenopausal women and

young women taking oral contraceptives (which raise

hormone levels artificially) were more sensitive to varia-

tions of cuteness than their respective comparison groups.

We suggest that cuteness sensitivity is modulated by female

reproductive hormones.

More than half a century ago, Konrad Lorenz proposed the

Kindchenschema as an innate releasing mechanism for care-

taking behavior and affective orientation toward infants, trig-

gered by features such as protruding cheeks, a large forehead,

and large eyes below the horizontal midline of the skull (Lorenz,

1943). Baby faces having these features are commonly de-

scribed as cute, and although cuteness has been shown to

modulate mother-infant interaction (Langlois, Ritter, Casey, &

Sawin, 1995), there are only a few psychophysical studies in-

vestigating responses of men and women to variations of phys-

ical properties of baby faces (Alley, 1981; Brooks & Hochberg,

1960; Gardner & Wallach, 1965; Hückstedt, 1965; Sternglanz,

Gray, & Murakami, 1977). Overall, findings from these studies

are not conclusive but suggest a possible difference between

men and women in perceiving cuteness. However, a question not

addressed in previous studies concerns what might underlie any

sex differences in adults’ ability to perceive infant cuteness.

An obvious cause might be differential interest in babies, but

there are other explanations. Given that Lorenz had conceived

the Kindchenschema as a biological mechanism, we decided

to investigate the possibility of a link to female reproductive

hormones.

STUDY 1

Study 1 aimed to explore the idea of a possible link between

cuteness perception and female reproductive hormones by

looking at groups of younger women, younger and older men, and

women aged slightly below and above the average age at

menopause in Britain.

Method

Participants

Twenty-four younger women (mean age 5 22.0 years, SD 5 1.8

years, range 5 19–26 years), 24 younger men (mean age 5 21.5

years, SD 5 1.5 years, range 5 19–26 years), and 24 older

women (mean age 5 53.2 years, SD 5 4.1 years, range 5 45–60

years) were investigated. Bearing in mind that reproductive

hormones might possibly modulate sensitivity to infant facial

cuteness, we used the older women’s median age of 52 years

(which is near the average age at menopause in Britain) to

subdivide the latter group into a group of women aged 51 years

and younger (mean age 5 49.8 years, SD 5 2.1 years, range 5

45–51 years), and a group of women aged 53 years and older

(mean age 5 56.6 years, SD 5 2.5 years, range 5 53–60 years).

No participant was taking hormone-replacement therapy or had

undergone hysterectomy. At a later stage of the study, a group of

11 older men (mean age 5 56.5 years, SD 5 2.3 years, range 5

53–60 years) was added to provide a point of comparison to

the women aged 53 through 60 and to the younger men.

Address correspondence to R. Sprengelmeyer, School of Psychology, University of St. Andrews, St. Andrews KY16 9JU, Scotland, e-mail: [email protected].

P S Y C H O L O G I C A L S C I E N C E

Volume 20—Number 2 149Copyright r 2009 Association for Psychological Science

The 48 younger participants reported not having children,

whereas all 35 older participants had children.

Cuteness Discrimination

Facial images of 130 female and 72 male Caucasian European

babies were collected. From these sets, we choose 28 pictures of

female and 30 pictures of male faces considered to be of a

sufficient quality for further computer processing. The female

baby faces were rated for cuteness on a 7-point scale (1 5 not

cute, 7 5 very cute) by 10 younger women (mean age 5 22.9

years, SD 5 3.6 years, range 5 20–30 years), 10 younger men

(mean age 5 24.2 years, SD 5 4.1 years, range 5 20–29 years),

and 10 older women (54.1 years, SD 5 4.1 years, range 5 45–59

years). The male baby faces were rated in the same way by

groups of 10 younger women (22.8 years, SD 5 3.3 years, range

5 20–29 years), 10 younger men (23.2 years, SD 5 3.7 years,

range 5 19–29 years), and 10 older women (54.0 years, SD 5

4.4 years, range 5 45–59 years). Ratings for female faces

correlated significantly between groups, Spearman’s r � .76, Cronbach’s a � .82. Similar results were obtained for the male faces, Spearman’s r � .73, Cronbach’s a � .82. This finding indicates that all participants had similar opinions concerning

what they considered cute and less cute baby faces.

On the basis of these ratings, we compiled four sets of pic-

tures: 10 female infant faces rated as cute (mean rating 5 4.4,

SD 5 0.8), 10 female infant faces rated as less cute (mean

rating 5 3.3, SD 5 0.6), 10 male infant faces rated as cute (mean

rating 5 4.8, SD 5 0.8), and 10 male infant faces rated as less

cute (mean rating 5 3.2, SD 5 0.7).

We took the pictures from these four sets and defined the

shape of each face with 174 manually marked facial landmark

points using the computer program Psychomorph (Burt &

Perrett, 1995). The average face shapes (Fig. 1) were then cal-

culated for each set. Next, we randomly chose five different faces

of female babies from the set of 28 pictures, and five different

faces of male babies from the set of 30 pictures. The average

a b

c d

e

+50% +25% 0% –25% –50%

Fig. 1. Creation of the experimental stimuli. First, four sets of baby faces were used to obtain av- eraged cute male (a) and female (c) baby faces and averaged less cute male (b) and female (d) baby faces. Next, these averages were used to apply the cuteness continuum to an individual baby’s face, as illustrated in (e). The cuteness continuum ranged from 150% (image altered to make the face more cute) to �50% (image altered to make the face less cute). The baby shown in (e) is 6 months old, and transformations were performed using the shape information from the faces shown in (a) and (b).

150 Volume 20—Number 2

Cuteness in Baby Faces

cuteness rating of these pictures was 4.5. The individual shapes

of these faces were also defined.

The photographs of the five female faces were shape-

transformed using the differences between the corresponding

landmark points of the cute and the less cute female averages. To

make the five individual female faces 25% or 50% cuter, we

added 25% or 50% of the differences to the values of the indi-

vidual’s face original landmark points. To make the individual

face 25% or 50% less cute, we subtracted 25% or 50% of the

differences from the individual’s face original landmark points.

The same procedure was applied to the photographs of the

male faces. This resulted in five pictures differing in cuteness

for each individual infant (50%, 25%, 0%, �25%, �50%). Details of the transformation procedure can be found elsewhere

(Tiddeman, Burt, & Perrett, 2001).

In each of the 200 pseudorandomized trials of the cuteness

task, two pictures of the same baby were shown, one each on the

left and right side of the screen. Participants had to decide which

of the two faces was cuter. In half of the trials, both faces were

presented in an upright orientation; in the remaining 100 trials,

faces were inverted. For each of the babies, there were two

upright and two inverted pairs of faces at each of five levels of

difference in cuteness (pairings used are in parentheses): 100%

(150% vs. �50%), 75% (150% vs. �25%, or 125% vs. �50%), 50% (150% vs. 0%, or 0% vs. �50%), 25% (150% vs. 125%, or �25% vs. �50%), and 0% (identical images). The 0% difference condition was included only to

balance the design, and was not used for analysis. The inverted

stimuli were intended to investigate the extent to which the

pattern of findings might be orientation-specific. However, they

led to near-chance performance in all conditions, and the data

for inverted faces are therefore not presented here.

Size Discrimination

A size-discrimination task was included for comparison to the

cuteness-discrimination data. In each of 50 trials, two squares of

different sizes were presented at random locations, one each on

the left and on the right side of the screen. A square with a side

length of 100 mm was paired with squares with side lengths of

99 to 95 mm. Participants had to point to the smaller square.

Results

We performed a repeated measures analysis of variance

(ANOVA) of the cuteness-discrimination data with task diffi-

culty (25%, 50%, 75%, 100% difference) as a within-subjects

factor and group (women aged 19–26 years, men aged 19–26

years, women aged 45–51 years, women aged 53–60 years, and

men aged 53–60 years) as a between-groups factor. The analysis

showed a significant effect of difficulty and group, Fs � 14.54, p < .001, as well as a significant interaction between these

variables, F(12, 234) 5 2.00, p < .05. As shown in Figure 2a, the

two groups of younger women were more accurate than the older

women and the two groups of men. This was true at all levels of

difficulty, but the group differences were greater at the easier

levels (i.e., when faces differed in cuteness by 75% and 100%).

To explore the group effect in more detail, Scheffé tests (p < .05)

were performed. These tests revealed no significant differences

among the following three groups: men aged 53 to 60 years,

women aged 53 to 60 years, and men aged 19 to 26 years. There

was also no difference in performance between women aged 19

to 26 years and women aged 45 to 51 years. However, the former

three groups differed significantly from the latter two groups.

To establish whether basic visuo-perceptual differences might

contribute to the observed pattern of results, we administered a

control task involving discriminating differences in the size of

two squares. Overall performance was 83% (SD 5 6.7%). An

ANOVA showed that performance declined with task difficulty,

F(3, 312) 5 85.40, p < .001, but neither the group effect nor the

Difficulty � Group interaction was significant, Fs � 1.53, p � .09. These results show that basic visuo-perceptual processing

did not differ among groups.

STUDY 2

In Study 1, we found a difference in cuteness discrimination

between women aged 45 to 51 and 53 to 60 years. Given an

average age at menopause of 51 years in Britain, these findings

suggest that reproductive hormone status might be linked to

cuteness discrimination. To explore this possibility, we looked at

cuteness and size discrimination across pre- and postmeno-

pausal women matched for age. In this study, only the upright

faces were presented. Otherwise, the procedures were identical

to Study 1.

Participants

This study included 20 healthy women, 10 of whom reported

having passed menopause at least 2 years before the study (mean

age 5 55.0 years, SD 5 1.9 years), and 10 who were pre-

menopausal (mean age 5 54.4 years, SD 5 1.8 years). No

participant was taking hormone replacement therapy or had

undergone a hysterectomy. The groups did not differ in age, mean

number of children, mean age of the youngest child, and mean

age of all children, ts � 1.13, p � .27, or in how much they liked children, Mann-Whitney z 5 �0.35, p 5 .73. The latter was rated on a 7-point scale (1 5 not at all, 7 5 very much). We also

asked whether participants had contact with children several

times per week, several times per month, or rarely. Most par-

ticipants reported contact with children several times per week.

No statistical group difference was found, w2(2, N 5 20) 5 1.29, p 5 .52.

Results

An ANOVA for the cuteness-discrimination data showed a sig-

nificant effect of difficulty, F(3, 54) 5 3.99, p < .05, and no

Volume 20—Number 2 151

R. Sprengelmeyer et al.

Difficulty � Group interaction, F(3, 54) 5 0.50, p 5 .68. A significant group effect, F(1, 18) 5 5.42, p < .05, indicated

reduced ability to discriminate cuteness in the postmenopausal

group (Fig. 2b).

An ANOVA showed that size discrimination declined with

task difficulty, F(3, 312) 5 85.40, p < .001, but neither the

group effect nor the Difficulty � Group interaction was signifi- cant, Fs � 0.84, p � .37. Overall performance was 86% (SD 5 6.0%).

STUDY 3

Results from Study 2 suggest the involvement of female repro-

ductive hormones in modulating cuteness sensitivity, with es-

trogen and progestogen as the most likely candidate hormones.

Therefore, we decided to address this possibility more directly

by comparing cuteness sensitivity in young women taking and

not taking oral contraceptives (combined pill containing estro-

gen and progestogen). Procedures used were identical to those in

Study 1.

Participants

We investigated 24 women, of whom 12 (mean age 5 21.4 years,

SD 5 0.5 years) were taking oral contraceptives (tested outside

the 7-day pill-free interval) and 12 (mean age 5 21.2 years,

SD 5 0.4 years) were not (tested at day 12.7 of the cycle, SD 5 7.6).

There was no significant age difference between groups, t < 1.

All participants reported not having children. Participants had

.4

.5

.6

.7

.8

.9

25 50 75 100 Difference in Cuteness (%)

P ro

po rt

io n

C or

re ct

Women Aged 19–26 Women Aged 45–51 Women Aged 53–60 Men Aged 19–26 Men Aged 53–60

.4

.5

.6

.7

.8

.9

25 50 75 100 Difference in Cuteness (%)

P ro

po rt

io n

C or

re ct

Premenopause Postmenopause

.4

.5

.6

.7

.8

.9

25 50 75 100 Difference in Cuteness (%)

P ro

po rt

io n

C or

re ct

Pill No Pill

a b

c

Fig. 2. Mean accuracy (� 1 SE) in the cuteness-discrimination task as a function of the difference in cuteness between the faces. Results are shown for (a) men and women of different age groups (Study 1), (b) age-matched pre- and postmenopausal women (Study 2), and (c) women taking and not taking oral con- traceptives (Study 3). Chance performance on this task was .5.

152 Volume 20—Number 2

Cuteness in Baby Faces

to rate how much they wished to have children of their own on a

7-point scale (1 5 not at all, 7 5 very strong). The groups did not

differ on this measure, Mann-Whitney z 5 0.59, p 5 .63. As in

Study 2, we also asked about the frequency of contact partici-

pants had with children; no group difference was found, w2(2, N 5 24) 5 0.00, p 5 1.

Results

The ANOVA for the cuteness-discrimination data showed a

significant effect of difficulty, F(3, 66) 5 17.58, p < .001; no

Difficulty � Group interaction, F(3, 66) 5 0.47, p 5 .70; and a significant group effect, F(1, 22) 5 5.10, p < .05. Women taking

oral contraceptives judged cuteness more accurately than wo-

men not taking oral contraceptives (Fig. 2c).

An ANOVA showed that size discrimination declined with

task difficulty, F(4, 88) 5 21.58, p < .001, but the group effect

and the Difficulty � Group interaction were nonsignificant, Fs� 0.35, ps� .56. Overall performance was 84% (SD 5 5.1%).

DISCUSSION

Study 1 showed that women aged 19 to 26 were more sensitive to

small variations of cuteness than same-aged men. Women aged

19 to 26 and 45 to 51 did not differ in performance, whereas

sensitivity to cuteness in women aged 53 to 60 was at the level of

the young men and a group of older men aged 53 to 60. Study 2

found differences in cuteness discrimination between age-

matched premenopausal and postmenopausal women, and

Study 3 revealed increased cuteness sensitivity in women with

artificially raised levels of progestogen and estrogen from the

contraceptive pill.

What might cause these differences? Study 1 could suggest a

cohort effect as a possible explanation for the difference be-

tween women aged 45 to 51 and 53 to 60, or alternatively, age-

related role transitions that might impact on experience with

infant faces and hence affect related perceptual skills. For these

reasons, the groups who participated in Studies 2 and 3 were

matched for age and for a range of variables related to experi-

ence with and fondness for children. Despite this, differences in

cuteness discrimination were found, ruling out cohort and social

role effects as likely causes.

We now turn to more biological explanations. The average age

of menopause in Britain (51 years) is near to the group separa-

tion criterion of 52 years used in Study 1. Findings from Study 1

therefore suggested that reproductive hormones could be im-

portant for processing cuteness. This idea would also explain the

difference between women aged 19 to 26 and men (whether aged

19–26 or 53–60). The possibility of a hormonal link to cuteness

discrimination was followed up in Study 2, where we found that

premenopausal women discriminated cuteness significantly

better than same-aged and similarly experienced postmeno-

pausal women. Levels of progestogen and estrogen both drop

sharply after menopause and are candidate hormones that may

mediate the group differences. We explored the potential role of

these hormones in Study 3 by comparing cuteness sensitivity in

young women who were and who were not taking the combined

oral contraceptive pill. Because women taking this particular

form of oral contraception have artificially raised levels of

estrogen and progestogen, we expected them to discriminate

infant cuteness better than women not taking the pill. Our

results fit this prediction.

At the moment, we do not know how hormones (directly or

indirectly) influence judgment of cuteness. However, basic

visuo-perceptual processing did not differ between groups,

making it more likely that hormones impact on postperceptual

processes.

Given that the concept of cuteness not only encompasses the

processing of specific physical properties of a baby face, but also

involves an affectionate, ‘‘heart warming’’ (Lorenz, 1943) ori-

entation toward the baby, it may be that hormones alter the

emotional response to a cute baby face. Neural substrates

involved in reward processing and maternal behavior are pos-

sible candidate structures linked to these processes (Panksepp,

1998).

We also have to consider the potential functional value

of being able to discriminate small variations in cuteness.

Although caring for a baby involves an investment of time and

resources, it is rewarding to be with cute babies. They are rated

as more friendly, cheerful, and likeable (Karraker & Stern,

1990), and a female’s willingness to adopt a baby (Volk &

Quinsey, 2002) and a mother’s sensitivity toward the baby

(Langlois et al., 1995) depend on its cuteness. Heightened

cuteness sensitivity may therefore help the mother to focus on

the newborn and modulate attachment.

Acknowledgments—We thank L. Ferrier, J. Stevenson, D.

Filippou, K. Dutton, A. Rodgrigues, and I. Stephen for help in

image collection and processing; C.J. Fiebach for help in pilot

work; G. Brown and M. Arnold for helpful discussions; and the

students and their families who entrusted us with their photos.

The study was supported by the Dietmar-Hopp Stiftung.

REFERENCES

Alley, T.R. (1981). Head shape and the perception of cuteness. Devel- opmental Psychology, 17, 650–654.

Brooks, V., & Hochberg, J. (1960). A psychophysical study of ‘cute-

ness.’ Perceptual and Motor Skills, 11, 205. Burt, D.M., & Perrett, D.I. (1995). Perception of age in adult Cauca-

sian male faces: Computer graphic manipulation of shape and

colour information. Proceedings of the Royal Society B: Biological Sciences, 259, 137–143.

Gardner, B.T., & Wallach, L. (1965). Shapes of figures identified as a

baby’s head. Perceptual and Motor Skills, 20, 135–142.

Volume 20—Number 2 153

R. Sprengelmeyer et al.

Hückstedt, B. (1965). Experimentelle Untersuchungen zum Kind-

chenschema. Zeitschrift für Experimentelle und Angewandte Psychologie, 12, 421–450.

Karraker, K.H., & Stern, M. (1990). Infant physical attractiveness and

facial expression: Effects on adult perceptions. Basic and Applied Social Psychology, 11, 371–385.

Langlois, J.H., Ritter, J.M., Casey, R.J., & Sawin, B. (1995). Infant

attractiveness predicts maternal behaviours and attitudes.

Developmental Psychology, 31, 464–472. Lorenz, K. (1943). Die angeborenen Formen möglicher Erfahrung.

Zeitschrift für Tierpsychologie, 5, 233–409. Panksepp, J. (1998). Affective neuroscience. New York: Oxford Uni-

versity Press.

Sternglanz, S.H., Gray, J.L., & Murakami, M. (1977). Adult prefer-

ences for infantile facial features: An ethological approach.

Animal Behaviour, 25, 108–115. Tiddeman, B.P., Burt, M., & Perrett, D.I. (2001). Prototyping and

transforming facial textures for perception research. IEEE Computer Graphics and Applications, 21, 42–50.

Volk, A.A., & Quinsey, V.L. (2002). The influence of infant facial cues

on adoption preferences. Human Nature, 13, 437–455.

(RECEIVED 1/24/08; REVISION ACCEPTED 7/30/08)

154 Volume 20—Number 2

Cuteness in Baby Faces