| Data collection method: |
Participants
Data collection targeted 5, 7, 9, 11, 13, and 17-year-olds. These six age groups attempted to capture children in the different phases of adrenal and gonadal hormone production: age 5 (all hormones at floor), age 7 (DHEA starting to rise), age 9 (increasing DHEA), age 11 (high DHEA, early gonadarche), age 13 (late gonadarche), and age 17 (adult). Children were recruited from primary and secondary schools in North-East England and a secondary school in South London (with comparable rates of deprivation); the 9-year-old sample was recruited from an ongoing longitudinal study in North-East England. Initially 401 participants were recruited; however data from 55 children were excluded due to noncompletion or evidence of bias in responding on the face task (more than 10 responses in a row using the same button), leaving 346 (176 male) participants in total. Participating children ranged from 4 to 17 years of age. With regard to participants’ ethnicity, 74.6% were Caucasian, 15.8% were Asian, 5.0% were Afro-Caribbean and 4.6% were ‘mixed’ or ‘other’. The children participating in the longitudinal study came from wide-ranging socioeconomic (SES) status backgrounds as indicated by maternal report on the Hollingshead Index (Hollingshead, 1975). The schools from which the other age groups were recruited similarly varied in SES as indicated by the percentage of children qualifying for free school meals. Active consent was obtained from parents/guardians and participants verbally consented on the day of the testing. The study was approved by the appropriate University Ethics Committee, and all participants were tested in accordance with APA ethical principles for research.
Materials and Methods
Procedure. All participants viewed the faces on computer screens. The 4- to 5-year-old participants completed the facial preference task verbally with an experimenter, either individually or in a closely supervised group of 2 or 3. Participants in all other age groups were tested in groups, which ranged in size from 6 participants for the younger ages (7 and 11 years) to 27 participants for the older ages. For all except the 9-year-old sample, data collection was conducted in the school’s computer room. Participants were seated at individual computer monitors that were screened at the side to prevent children viewing each other’s responses. The 9-year-olds visited the University’s developmental laboratory with their parents, and completed the task alone at a computer. Children aged 5 to 7 years completed the 80 trial face test in one sitting with short breaks as needed. Children received a reward of a sticker for each 20 trials that they completed. The 9-year-olds completed the 80 trials in two equal blocks, with a break in between during which they completed a questionnaire on their family (note that these questionnaire data are not reported here). Participants ranging in age from 11 to 17 years completed half of the 80 trials, followed by the Pubertal Development Scale (see below), and then the remaining 40 trials of the face preference test. For all age groups, the procedure lasted between 15 and 20 minutes.
Stimulus format. Although using ‘masked’ images (in which hair and clothes are blanked out, leaving only the face) are commonly used in research with adults (for discussion see e.g., De Bruine, Jones, Smith, & Little, 2010), children might find masked images odd and thus struggle to judge them effectively. In order to establish whether masked or unmasked images were more appropriate for the age range of children in the present study, a pilot study involving 22 individuals (14 boys) ranging from 6 to 16 years of age was conducted. Participants were asked to rate whether they preferred the high- or low-trait versions of faces which varied on averageness or femininity (as described below) for both masked and unmasked faces. The masked and unmasked faces did not differ on mean preference scores for averageness, t(21) = 0.52, p = 0.61, d = 0.23, or femininity, t(21) = 0.63, p =0.54, d = 0.27. We therefore decided to use unmasked versions of all stimuli in the main study due to their greater familiarity and ecological validity.
Attractiveness stimuli. Front-facing, non-smiling images of eight male and eight female individuals, selected for differing attractiveness from previous research (Boothroyd et al., 2008) and the CALPAL database (http://vitallongevity.utdallas.edu/facedb/), were paired so that each pair included an image of an attractive individual and an unattractive individual of the same sex. There were four pairs of female images and four pairs of male images. The relative attractiveness of the resulting eight pairs of images was pre-rated by 104 adults who indicated for each pair which face they found more attractive. Across all adult raters the putatively more attractive stimuli were selected on 94.8% of trials, showing strong validity of the face pairings.
Femininity, health, averageness and symmetry stimuli. Four male and four female computer-generated composite faces were each created from six individual faces of undergraduate students using the program Psychomorph (Rowland & Perrett, 1995; Tiddeman, Burt, & Perrett, 2001). Following Boothroyd et al. (2009), these original composite faces were then manipulated to create the test stimuli. Briefly, the composite face was manipulated along the given dimension (symmetry, averageness, femininity, health) in both a positive and negative direction to create two new faces. Thus, for each of the four dimensions, there were eight male and eight female faces.
Facial Preference Task. Participants were presented with a pair of faces on a computer screen and asked to select which face they found more attractive. The 5- to 9-year-olds were requested to “Choose the face you like better”. The experimenter read this question out to the 5-year-olds, who indicated their preference verbally. The 11- to 17-year-olds were asked to “Choose the face you find more attractive”. All participants selected their chosen face using an extended forced choice paradigm. Having chosen which face was more attractive, participants then rated the degree to which they preferred the face using a scale at the bottom of the computer screen. Participants could indicate that they liked their chosen face “a lot more”, “more”, “a bit more”, or “a little bit more”. These responses were read out by the experimenter to the 5-year-olds, who then indicated verbally how much they preferred the chosen face. Participants viewed each pair of faces twice, giving eight male and eight female trials per trait. Each participant was given a score indicating the proportion of trials in which the more attractive/symmetric/average/feminine/healthy face was selected.
Pubertal development. The Pubertal Development Scale (PDS; Petersen, Crockett, Richards, & Boxer, 1988) was completed using maternal-report for the 9-year-old participants, and self-report for those aged 10 to 17 years. Petersen et al. (1988) found that maternal-report was most similar to physician ratings of pubertal development prior to age 11, whereas self-report was more accurate in older children. The PDS asks respondents to estimate the extent to which they (or their child) have developed on five key physical traits which are directly or indirectly affected by adrenarche or gonadarche: height, skin changes, pubic hair, breast development and menarche (for girls), and facial hair and voice breaking (for boys). For each physical trait, response options were: not yet started (1 point), barely started (2 points), definitely started (3 points), seems complete (4 points). Scores range from 5 for those who have not experienced any changes, to 20 for those who have completed all changes. Participants were coded into pubertal stages (pre-, early-, mid-, late-, and post-pubertal) on the basis of their scores. For boys, pubertal category was assigned from the scores on the body hair growth, voice change, and facial hair growth: 3 = pre-pubertal; 4 or 5 (with no 3-point responses) = early-pubertal; 6–8 (with no 4-point responses) = mid-pubertal; 9–11 = late-pubertal; 12 = post-pubertal. For girls, pubertal category was assigned from scores on body hair growth, breast development, and menarche: 2 and no menarche = pre-pubertal; 3 and no menarche = early-pubertal, >3 and no menarche = mid-pubertal; </=7 and menarche = late-puberty; 8–12 and menarche = post-pubertal. |
