This chapter aimed to discuss the main findings of this study which were gender difference in total face recognition, d-prime total face recognition, same-sex face recognition and opposite-sex face recognition. The results of this study will be evaluated with regard to some possible limitations. Recommendations for future research and conclusions were included as well.
Gender differences in total face recognition
The first research question of this study is to investigate total face recognition score between male and female. Total face recognition score is measure by percent of correct face recognition by participants that included number of hits and correct rejections out of 60 faces in the recognition task. The results of the data analyses suggest that there was a significant difference between male and female in total face recognition, t(48) = 3.375, p<0.05. Also, female participants (M=65.665, SD=6.702) scored higher than male participants (M=58.534, SD=8.169) in face recognition task.
Past research has found that females tend to recognize more faces than males in recognition task (Rehnman & Herlitz, 2006; Herlitz & Rehnman, 2008). The result obtained from this study was consistent with the prior research finding. Thus, the research hypothesis that female would recognize more faces than male was fully supported by the past research findings.
This result replicates the findings of Rehnman and Herlitz (2006) that aimed at understanding sex differences in face recognition. One explanation on the findings was superior female face recognition ability may be a function of the greater exposure of females to female face through the media (Loftus, Banaji, Schooler & Foster, 1987 as cited in Clifford & Bull, 1978). Besides, in the view of societal influences, Cross, Cross and Daly (1971) considered the role of magazine in female face recognition performance (as cited in Winters, 2009). Nowadays, the media and fashion magazine bombarded women with images of female models and actresses who look beautiful which could increase female exposure to various faces. Consequently, this would be providing them with greater opportunity to learn encoding strategies (Loftus et al., 1987 as cited in Clifford & Bull, 1978).
In addition, Clifford and Bull (1978) proposed that females greater in social attentiveness could have an effect on their face recognition superiority (Loftus et al., 1987). Past research have showed that males preferred “object” oriented jobs, while females preferred “people” oriented jobs (Winters, 2009 as cited in Lippa 1998). Hence, females tend to seek out more people-oriented jobs due to their social interest predisposition. In these job circumstances, they could have a great exposure by meeting up new people. Moreover, they need to socialize and strengthen their interaction and relationship with other people (Herlitz & Rehnman, 2008). Therefore, these differences between males and females in social preferences could promote females attention to various faces, as well as contributing to gender differences in face recognition.
Furthermore, another possible explanation on this finding was females could perform better than males in episodic memory task (Association for Psychological Science [APS], 2008). When the episodic memory tasks requiring verbal processing, result showed that females tend to perform better than male. For examples, tasks of remembering words, objects, pictures or everyday events. However, it also indicated that men outperformed women in visuospatial episodic memory tasks. Besides, Herlitz and Rehnman (2008) also revealed that women performed better than men in tasks requiring little to no verbal processing, such as recognition of human faces. They also concluded that episodic memory advantage favoring females when they utilize verbal abilities and decreases when visuospatial abilities are required (Herlitz & Rehnman, 2008).
Gender differences in d-prime total recognition score
This research finding showed that there was a significant difference between sex and d-prime total recognition score, t(48) = 3.592, p<0.05. Female participants (M=0.965, SD=0.433) had higher score in d-prime total face recognition compared to male participants (M=0.499, SD=0.484). The d-prime total face recognition score of participants is computed by the ratio of hits to false alarms. D-prime is a discriminability or sensitivity index derived from signal detection theory (Stanislaw & Todorov, 1999), typically using in Yes/No recognition task to calculate the score. It was calculated by determining the different between z score of hit and false alarm. In contrast, total face recognition variable from research question 1 only computed the percent correct recognition out of 60 faces. The number of false alarm and miss rate were excluded.
Past researches have indicated the significant gender differences in d-prime face recognition score. Thus, the finding of prior studies was replicated in this study. The explanation for this finding could be as similar as the explanation of finding of sex differences in total face recognition. For examples, females exposure to female face through the media and magazine, females social attentiveness and preferences effect (Loftus, et al., 1987 as cited in Clifford & Bull, 1978; Cross et al., 1971 as cited in Winters, 2009), and females advantage in episodic memory task (APS, 2008), could be the factors of superior female face recognition ability.
Gender differences in same-sex face recognition
This study hypothesized that females would recognize more same-sex faces than males in face recognition task. The hypothesis was supported by the finding of this study. The result indicated that there is a significant gender difference in total same-sex face recognition, t(48) = 2.938, p<0.05. Females significantly scored higher in total same-sex face recognition (M=8.520, SD=2.064) than male participants (M=6.600, SD=2.533).
Previous researches suggested that females could be more precisely identify females’ faces than male do (Lewin & Herlitz, 2002; Rehnman 2007). In the same time, there were no studies proving a male advantage in remembering female or male faces have been constituted. The purpose of this research finding was to examine the same-sex bias effect in face recognition and it produced a consistent result with the earlier finding. Past research proposed that females higher face recognition ability to same-sex faces was magnified by their own-sex bias (Rehnman 2007).
The research done by Rehnman (2007) also discussed female own-sex bias in term of differences in interest in faces. Besides, some studies showed that human interest could have influencing the episodic memory performance (Rehnman & Herlitz, 2006 as cited in McGivern, Huston, Byrd, King, Siegle, & Reilly, 1997; McKelvie et al., 1993). Females could perform better than males when the material consists of female-oriented objects which could catch their attention. Interestingly, some researches have shown that human faces are considered as female-oriented objects that could grab interest of female (Rehnman & Herlitz, 2006).
A study on infants’ behavior and development proposed that biological origin may also play a role in female face recognition advantage. The research done by Connellan, Baron-Cohen, Wheelright, Batki, & Ahluwalia (2000) found that female infants showed a stronger interest in the face than male which could evidently establish that biological origin contributed to face recognition. Although female own-sex bias to female faces is reported, there was no reliable own-sex bias showed by males in face recognition (Lewin & Herlitz, 2002). The study has reported there was no significant difference on both female and male faces recognized by male participants.
Gender differences in opposite-sex face recognition
In the study of gender differences in same-sex face recognition, it was also important to consider opposite-sex face recognition. This finding showed that there is no significant difference between males and females in opposite-sex face recognition, t(48) = 3.592, p>0.05. Both males and females did not show superior ability in the recognition of opposite-sex faces.
Though, previous research reported that males recognized more female faces than male faces (Winters, 2009). Winters (2009) explained this opposite-sex bias as males’ interest on female face could increase their mating opportunities. According to theory formulated by Buss (2007), the likelihood of mating opportunities could be increased when males recognizing female faces better (Winters, 2009), which in turn could improve the relationship of both parties. However, some studies have suggested males recognized both female and male faces equally well (Rehnman, 2007; Lewin & Herlitz, 2002).
Limitations of study
This study made an important contribution on the findings of gender differences in face recognition, same-sex face recognition and opposite-sex recognition. However, there were some limitations that being found in this study.
The first limitation of this study was the number of participants in this study. This study only recruited 50 participants in the experiment which could generate the result that might not be able to generalize to the whole male and female population. Besides, the external validity of this research is not well controlled due to lack of participants with different cultural and racial backgrounds. There were only 2 Indians took part in this research as compared to 48 Chinese. The result might not be able to generalize to multicultural society in Malaysia. In addition, the age range of the participants were between 21s to 23s years (M = 22.04, SD = 0.57) was too small to be generalize to all people with different age.
Next, the second limitation of this study was the response bias which was not computed in this study. Other than d-prime (d’), response bias (c) also derived from the signal detection theory. Response bias is used to compute the score when the task required subjects to determine ambiguous and uncertain stimuli which can be obtained either by a known process (signal) or by chance (noise) (Abdi, 2009). In the same time, it is also examined the extent between two response in which is more likely to happen than another. For example, the participants might be more likely to answer “No” rather than “Yes” in an ambiguous condition. Consequently, result could show more “No” response than “Yes”. Thus, response bias score could take the number of hits (correct “Yes”), false alarm (incorrect “No”), correct rejection (correct “No”) and miss rate (incorrect “No”) into the account (Abdi, 2009).
The third limitation of this study was other ethnicity and race bias that might affect the performance of the participants. In this experiment, the faces used in recognition task were mostly White and Black people from western country rather than Asian or Malaysian itself. It was assumed that lack of other-ethnicity knowledge and low level of face familiarity could reduce the face recognition ability of the people with different backgrounds.
Furthermore, the forth limitation could be also the reliability and validity of this modified experiment. The later face recognition task in this experiment was adjusted due to lack of resources and software. During the face recognition task, the faces were shown in list in several pieces of paper rather than using computer. The participants were told to response by answering “Yes” or “No” on the paper. The low-definition of faces that showing in paper might influence the performance of the participants. Though, the pre-testing was conducted before carrying out the experiment.
Lastly, the fifth limitation was there were some famous people faces taken from internet due to lack of appropriate face image resources, such as artist and sportsperson. The face images set provided from the database were not enough to use in this 60 faces experiment. Some faces were not suitable because of facial hair and not showing neutral expression. Therefore, a small amount of faces was taken from the internet that might be seen by the participants before the experiment. This might affect the outcome of the study.
Recommendations of study
This study revealed that there was a significant different between male and female in total face recognition, d-prime total face recognition and total same-sex face recognition. However, the study found no significant gender differences in opposite-face recognition.