Chapter 5: Discussion, Limitations, and Conclusion
5.1. Discussion of Results
5.1.3. Additional Analysis
According to previous research, it has been stated that regarding SS, HeM would be most alike to HoW in their preference and response (Lubke, et al., 2012; Berglund, et al., 2006) and that HoM would be most alike to HeW (Savic et al., 2005). The outcomes of this study give credence to the findings produced by those aformentioned studies, as rankings of SA were significantly concordant between HoM and HeW as well as HeM and HoW, this result was also found in rankings of SM suggesting that these groups ranked the attractiveness and masculinity of the six t-shirts similarly.
For measures relating to VS all combinations of sex by SO groups were significantly concordant. This may be due to, as mentioned previously, societal mores of visual attractiveness and masculinity.
5.1.3. Additional Analysis
such as T in men (Bao & Swaab, 2011). However, if the neuro-hormonal hypothesis is correct, then brain differentiation of homosexual individuals would be similar to heterosexual opposite sex individuals. Therefore, behaviours in homosexual individuals may be induced by both sexually dimorphic hormones specific to sex and brain differentiation similar to heterosexual opposite sex individuals, thus exhibiting traits typical of both sexes.
Heterosexual individuals on the other hand, are associated with the undifferentiated classification of the PAQ, and thus endorse traits typical of neither sex. This may perhaps be due to the change in gender roles, which has been discussed previously, for example the introduction of feminism and post-modernism. This result may also perhaps be due to the considerable cultural variations in the South African context from which the sample was drawn, which may differ regarding gender norms as the sample included individuals from multiple cultural backgrounds. According to Helman and Ratele (2016), who conducted a study investigating the construction of gender norms in various South African families, norms of gender inequality are largely influenced by culture, socio-economic status and religion. They noted that a majority of families produced a somewhat patriarchal discourse when interviewed, however, the different social classes in terms of culture and socio- economic status did show a difference in their construction of patriarchal norms. Therefore,
“the South African context should…be viewed as ‘a patchwork quilt of patriarchies’ in which equitable gender relations extend beyond the social categories of race, class and culture”
(Helman & Ratele, 2016, p.2). Helman and Ratele (2016) found that gender constructions in a South African context, although predominantly patriarchal, have the propensity to be fluid.
The researchers suggested that rather than defining families as either egalitarian or traditional, that they fall somewhere on a continuum of these ideals (Helman & Ratele, 2016).
Race and Age
Although this study did not aim to investigate whether age or race affected judgements of attraction or masculinity, 𝑊̃ tests were run to investigate whether age and race were concordant as these are recognised as potential extraneous variables.
It has been suggested that race is associated with culture and culture with food, thus affecting the smell of BO produced by individuals (van Beek, 1992). Therefore, to reduce the effect of culturally dependant BO, only two race groups (Black and White) were recruited to produce the stimulus. However, all races were included as judges. This additional analyss aimed to
assess whether any of the races were more concordant. According to the results produced.
Black individuals showed no significant agreement for any of the ranking conditions, suggesting that this race would not have been a confounding variable in any of the other analyses and that the stimuli were not concordant among the Black population. White individuals were only concordant regarding the VS. This may be explained by a social bias as blatant indications of status are visible in the images of the male stimulus participants, thus suggesting that White individuals are more biased by social mores when judging VS.
Indian participants were concordant in all ranking conditions except for SM and Coloured individuals showed significant concordance for all ranking conditions. This suggests that perhaps Indian and Coloured individuals may have confounded the analyses. Interestingly, the race groups that could potentially have produced the greatest confounding effect (Indian and Coloured) were not included in the recruitment of the stimuli participants. This leaves scope for future studies investigating whether different race groups can discern other races through smell.
According to Jankowiak et al. (1992), attractiveness judgements are commonly influenced by the age of the individual being judged. It is for this reason that all the male stimuli participants were within a similar age range. Furthermore, according to Lippa (2007) both HeM and HoM regard age as a factor which influences attractiveness. Therefore, it was important in this study, to ascertain whether there was concordance amongst the ages. Both younger age categories were concordant in their rankings regardless of the ranking condition.
The first of the two younger age categories consisted of respondents between the ages of 18 and 25, and the other group ranged from 26 to 30 years of age. The oldest age group however, was left out of the analysis as there was only one individual older than 30 years. As all the age categories were significant regardless of ranking condition, age may be considered a confounding variable as is suggested by Lippa (2007) and Jankowiak et al. (1992).
A comment on the quantitative procedures used
In addition to the discussion of the results it is also pertinent to discuss the uniqueness of the analytical approach taken in this study, as statistical procedures such as Kendall’s 𝑊̃ and bootstrapping are not commonly implemented in social science research designs. However, they do provide a valuable alternative to the more traditional parametric statistics used when assumptions are violated and sample sizes small. Kendall’s 𝑊̃ provides a measure of inter-
judge reliability that covers a wider range of judgments than means or medians (Siegel, 1956). This measure provides a simple and time efficient way of calculating over-all agreement amongst k sets of rankings (Siegel, 1956). According to Sheskin (2007), 𝑊̃ provides a measure for “data that are rank-ordered by more than two judges” (p.1388), alternatively one could use Spearman’s rho as “𝑊̃ for [k] sets of ranks is linearly related to the average value of Spearman’s rho which can be computed for all possible pairs of ranks”
(p.1387). Computing multiple calculations for all possible pairs of rankings to find the average agreement amongst a large sample of k judges, however would be time consuming and likely increase the family-wise error rate (Siegel, 1956). Therefore, 𝑊̃ was a more suitable measure to find agreement amongst the ranks given by the judging participants.
Furthermore, as the data collection procedure required participants to rank order the stimuli, a parametric factorial method would not have been appropriate. Bootstrapping, provides another unique empirical method for analysing the data, which enables the researchers to predict and estimate population parameters, without making parametric assumptions (Chernick, 2008;Winston, 2004; Sprent, 1989). Furthermore, with bootstrapping, one can estimate population confidence intervals using the percentile method which although could be refined with the use of a bias corrected model with an acceleration constant, does provide good approximations of the 95% confidence interval parameter.
5.2. Limitations, Recommendations and Conclusions