Homophily Attractiveness for Women in Online Dating in Taiwan

Online dating is a significant and worldwide social phenomenon. Whilst research in western countries has provided valuable insights on how users perceive each another when viewing their profiles, little research has been done in Asia, where culture and cultural environments are different. In order to evaluate homophily, in an experimental approach, we evaluated what characteristics single women seek in potential romantic partners. We interviewed 55 young single women who are engaged in online dating. We asked participants to describe themselves in terms of personality, habits, family orientation, and etcetera. In addition, using the same variables, we asked them to describe their ideal partners in a second questionnaire. Women exhibited a low tendency of homophily towards their desired partners, seeking men who are not similar to themselves in life course attributes.


Introduction
As a social species, humans have a strong tendency to develop relationships with romantic partners during their lifetimes.As a result of modern technological innovations, online solutions that allow people to meet potential partners online in numbers and ways that were impossible decades ago are widely available today.Computers and internet have changed our lives.They have transformed how we dream, flirt with people, fall in or out of love, and marry.In short, technology has become a part of our modern way of loving.
The great number of dating websites or applications offer an efficient path for online users to meet new people and develop different kinds of relationships (Hardey, 2004;Ellison, Heino & Gibbs, 2006;Gibbs, Ellison, & Heino, 2006).Furthermore, they offer important advantages over traditional dating such as larger databases of potential mates and wider geographical locations and or different cultural backgrounds.Such websites or applications offer low threatening environments and contexts for new users or shy people to build relationships with potential mates.We can choose from those databases people who we believe can be the best relationship partners and skip those who we think might not (Finkel, Eastwick, Karney, Reis, & Sprecher, 2012).
Dating websites and applications allow users to filter potential romantic partners based on location and other demographic characteristics such as, age, race, religion, education and hobbies.Previous research has shown the individual propensity of people to search potential partners who have similar individual characteristics.The reason is that the actual and perceived similarity between two potential romantic partners is positively correlated with mutual attraction and with subsequent relationship satisfaction (Fiore & Donath, 2005;Blossfeld, 2009;Skopek, Schulz, & Blossfeld, 2011).
People interested in finding potential partners with similar characteristics seek clues in online profiles in order to find users who might be similar.In this context, users build their profiles in an appealing way because they are interested in presenting themselves in a desirable and realistic way (Toma, Hancock, & Ellison, 2008;Whitty, 2008).In addition, their profiles are designed in a way that is marketable in specific cultural contexts (Jagger, 2005).However, the information users post in their profiles often differ from the information that can be gathered in face-to-face interaction (Fiore & Donath, 2005).

Technology Acceptance Model (TAM)
Consumers may use high-technology products in order to get desired benefits of the technology whilst enjoying the experience of using them (Mick & Fournier, 1998).The Technology Acceptance Model (TAM) was adapted from the theory of reasoned action developed by Fishbein and Ajzen (1975).They proposed that a person's behavioural intention to adopt a particular piece of technology is determined by the individual's attitude toward the use of the technology.In turn, attitude is determined by two beliefs of the person: "perceived usefulness" and "perceived ease of use" (Kulviwat, Brunner II, Kumar, Nasco, & Clark, 2007).Mick and Fournier (1998) argued that technology may trigger in consumers two kinds of opposite feelings: positive and negative.Regarding the positive feeling, consumers can experience a pleasant interaction, excitement, and confidence when considering adopting the technology.The negative feeling is concerned with consumers getting annoyed, worried, or scared.All these emotions are relevant when adopting and using high-tech innovations.

Internet Disruption in the Formation of Human Relationships and Industry Convergence
The Internet allows access to a great number and diversity of potential partners, most of whom would be unknown or inaccessible without it.Before the creation of the Internet, Kerckhoff (1964) proposed the "Field of Eligible" or space where people can choose potential mates, was limited to an individual's immediate social network.Online dating represents an important unprecedented access to potential partners, increasing the "Field of Elegible" (Kerckhoff, 1964).Furthermore, individuals who adopt and use technological platforms, after meeting online, can thereafter move offline (Ellison et al., 2006).
Online dating platforms are qualitatively different from many other online settings due to consumer's anticipation of face-to-face interaction and due to the fact that social interactions and practices are in constant development (Parks & Floyd, 1996;McLaughlin, Osbourne, & Ellison1997;Roberts & Parks, 1999;Utz, 2000;Bargh, McKenna, & Fitzsimons, 2002;Ellison et al.,2006;Gibbs et al.,2006).
Today, a number of factors have made online dating a more viable option.One of the most important of these factors is easy access to the Internet from computers or hand-held devices.In addition, the affordable cost of online dating platforms plus the greatly reduced social stigma associated in the past with online dating have all contributed to the consumer's perception (perceived usefulness) of online dating as a viable and efficient way to date or find long-term relationship partners (St. John, 2002).Internet dating platforms offer sophisticated self-presentation options, and they use algorithms to match users with suitable potential partners in databases with millions of users ( (Ellison et al., 2006;Finkel et al., 2012).

Theory of Social Information Processing
When users adopt a technology which allows them to meet potential partners online, they form impressions of each other, which helps them decide whether or not to pursue a relationship (Lampe, Ellison, & Steinfield, 2007).In face-to-face meetings, people form initial impressions, in about three minutes (Sunnafrank & Ramirez, 2004).These impressions are managed strategically, emphasizing some characteristics whilst de-emphasizing others.These processes of impression formation and management also happen online, but in a different way due to constraints of CMC; for example it's not possible to know the linguistic accent or dressing style of someone behind a dating user profile (Lampe et al., 2007).According to the theory of Social Information Processing, people would make inferences about others using any available cues, such as spelling ability (Walther, 1992;1996;Tidwell & Walther, 2002).
The notion of "hyperpersonal interaction" (Walther, 1996), suggests that since users communicate online and because cues from other users are few or arrive slowly, they might develop higher levels of affinity for each other than they would face-to-face.This could be due to users' tendency to fill in the information blanks optimistically if they lack sufficient information about another user (Walther, 1996).For example, Hancock, Toma & Ellison (2007), in a laboratory study with college students, found that students liked each other more when they met in text-only chat and less when they met face-to-face.

Mate Selection as Social Exchange
Based on marriage market findings and social exchange theory, it can be assumed that users seeking for partners in dating websites or apps try to increase their "expected subjective utility" by exchanging resources socially with other users (Skopek et al. 2011).Given the limited and incomplete information in online profiles, users set up a minimum standard of the expected suitable match they can accept (Todd & Miller, 1999).The minimum standard is based on the users' Background and Resources and is built upon the users' own perceived mate value.Therefore, if users regard themselves as attractive and desirable, then their aspirations will be higher, and their minimum standard will narrow down the number of acceptable partners (Todd & Miller, 1999).
As previously proposed by researchers, educational similarity is quite important for romantic relationship development.When applied to online dating, educational similarity makes it easier to find a common lifestyle because it's associated with similar cultural interests that decrease struggles in communication and help develop romantic relationships (DiMaggio & Mohr, 1985;Kalmijn, 1994;Kalmijn & Bernasco, 2001).
Previous research tends to consider only resources (including education as a resource) as important factors for relationship formation (Becker, 1973;1974;Anderson & Hamori, 2000;Skopek et al. 2011).In the present study, for practical purposes, we make a distinction between background and resource similarity.In a person's Background we include the following factors:  ethnicity  sociocultural aspects  education  marital status  values  personality (influenced by cultural and family environments) In a person's Resources, we include the following factors:  physical attributes  perceived attractiveness seen through pictures  kind of job  money income Users tend to seek a cost-benefit balance in potential relationships.For example, the higher the resource value of a user, the higher will be the value of the resources demanded from a potential partner.Users stop their search for potential matches when an alternative that exceeds their minimum standard is found and then assess it (Todd & Miller, 1999).When a user finds a potential partner which meets his or her minimum standard, then they will begin a relationship, providing both the user and his or her potential match believe that there will be potential benefits.However, if the potential match fits neither the minimum standard nor brings potential benefits to a relationship, then it can be expected that users will keep searching until both requirements are met (see Fig. 1).

Homophily in Dating
Researchers found that users change their presentation of themselves and the characteristics of their preferred partner depending on the kind of relationship they seek.Users seeking for long-term relationships would mention their personal competences, their wish for courtesy and honesty.Furthermore, these users displayed an active or cultured lifestyle.Alternatively, users seeking short-term intimate encounters would mention their physical characteristics or sexual skills.Researchers found that both perceived and real similarity between users correlates with attraction, which can later turn into relationship satisfaction (Fisher 1992;Brehm, Miller, Perlman, & Campbell, 2002;Fiore & Donath, 2005;Hitsch, Hortaçsu, & Ariely, 2010).This homophily can be exhibited by users to differing degrees for different characteristics.
Previous research on dyadic interactions showed that users sought people similar to themselves much more often than chance would predict.This trend is similar to what would be expected in the offline world.A study by Fiore and Donath (2005) found that the preferences of users for certain attributes of life course were strongly same-seeking, such as the following:  the desire to have children In addition, users who already had children searched for users with a similar number of children.
Research in the marriage market by Becker (1973;1974), led him to propose that "like is attracted to like" through a hypothesis of positive sorting by non-market traits.According to Becker, when a male and a female share personal attributes to the maximum possibility, then the benefits of getting married compared to staying alone are higher.Therefore, we can assume that users of dating websites and apps will initiate contact or reply to other users if both users have similar perceived resources and backgrounds, otherwise they would avoid contact and keep searching.If a user accepts another user who has lower resources and different background, the cost-benefit ratio will be negative.Given the great number of online daters, there is competition to attract the desired potential couple, therefore we can assume that online dating users seek couples with similar backgrounds and resources.Skopek et al. (2009;2011) proposed that this mechanism would be symmetrical and independent of gender.They also suggested that the higher the resources of the user, the stronger would be the tendency.

Hypothesis
To assess homophily, we decided to separate Background homophily from Resources homophily.To test the former, we included some aspects of personality and values such as:  Being easy going  Risk taking Earlier research has shown repeatedly that similarity positively facilitates future romantic relationships, mating, and marrying (Brehm et al. 2002, Fisher, 1992, Kalmijn, 1998;Fiore & Donath, 2005;Blossfeld, 2009).We tested homophily in Taiwanese women seeking romantic partners.We evaluated the types of partner that women are interested in finding.Relying on Becker (1973), we expect that women will seek potential couples that have similar perceived backgrounds.We worked with null hypotheses, and hypothesized that:  Null hypothesis 1: Women seek men who are similar in terms of being easy going  Null hypothesis 2: Women seek men who have similar risk tendencies  Null hypothesis 3: Women seek men who have a similar degree of self-control  Null hypothesis 4: Women seek men who have a similar life-plan  Null hypothesis 5: Women seek men who give the same value to money  Null hypothesis 6: Women seek men who have similar family orientation

Methodology
To test our homophily hypotheses, this study used empirical data gathered from questionnaires distributed to 55 single women aged between 21 and 25, who regularly used online dating websites to find romantic matches, at least twice a week.All females were college and master students in Feng Chia University.Fiore and Donath (2005) to analyse homophily in their study, used different profile characteristics that users could specify about themselves and about the partners they would like to meet:  Age  Importance of height in a partner Fiore and Donath (2005) considered some characteristics more bounding than others, meaning users were more likely to seek someone similar to themselves on that dimension, for example, smokers seeking other smokers.To determine the bounding strength of descriptors in their data, they compared the percentage of contacts between two users who shared the same value for a characteristic with the percentage of contacts normally expected to share the value if one male user and one female user from the active user population were paired randomly.
Since we targeted only women, our study differed slightly from Fiore and Donath (2005).We developed a questionnaire including some of the variables studied by Fiore and Donath (2005), while adding new ones adapted to Taiwan's cultural environment.Our questionnaire was designed with the aim of discovering what single young women want and how much similarity they expect in the ideal romantic partner in their age group.Based on the selected variables, we asked the participants to describe themselves (see Table 1).After they finished the questionnaire, we asked them to answer a second questionnaire using the same variables to describe their ideal partners.The questionnaires were posted in Google Docs, and were posted for a period of one week.We normalised the data using logistic regression.Then we performed Pearson correlations between the following variables:  Personality of females and expected personality of their desired matches  Risk taking of females and expected risk taking of their desired matches  Self-control of females and expected self-control of their desired matches  Life-plan of females and expected life-plan of their desired matches  Importance of money for females and expected importance of money for their desired matches  Family orientation of females and expected family orientation of their desired matches Thereafter, we performed Paired t-Tests assuming unequal variances to determine the relationship between the same variables.We used six Null hypotheses for each t-Test: H 0 : µ 1 = µ 2 ("the paired population means are equal") We performed our analysis using SPSS 22.

Results
We gathered data for 55 single young women studying in our university.Women described themselves as average in their personalities with a positive attitude (Table 2).The correlation shows that the personality of females and that of their desired matches are weakly positively correlated (Table 3).Further, the two sets of data differ significantly, t(54)=2.461,p=0.017, therefore, we reject Null Hypothesis 1. Women seek men who are more easy going than themselves, which is illustrated in Table 2. (2-tailed) .009N 55 55 **.Correlation is significant at the 0.01 level (2-tailed).
Women described themselves as risk-takers to a certain extent and appeared to prefer partners who liked risk-taking as much as they did (Table 4).We found a significant correlation between both sets of variables (Table 5).However, the two sets of data differ significantly, t(54)=2.142,p=0.037.Therefore we reject Null hypothesis 2 (Women seek men are willing to take more risks than themselves).Females tended to describe themselves as people who sometimes can control themselves (Table 6).Despite only a moderate correlation being found for self-control (Table 7), the two sets of data differed significantly t(54)= 5.190, p=0.00003.Therefore we reject the Null hypothesis 3. We found that women seek men who exhibited a much higher degree of self-control than themselves.Regarding Life-plan, few women classified themselves as willing to open a company.Most of them chose to have a small business or an ideal job (Table 8).Only a weak correlation was found between women and their desired couples (Table 9).However, the two sets of data are different significantly t(54)=2.784,p=0.007.Therefore we reject Null hypothesis 4. Women will seek men who have better life-plans than themselves, this can be appreciated in Table 11, women rather couples that are more willing to open their own companies or have small businesses.Most women stated that money is somewhat important for their future lives but not the most important thing (Table 10).A significant correlation was found for money importance (Table 11), Further, the two sets of data are not different significantly t(54)=-0.971,p=0.336, therefore we accept Null hypothesis 5. Women will seek men who give the same value to money.Most women said that they wanted children or that they may want them (Table 12).There was a significant correlation found for family orientation (Table 13).However the two sets of data are different from each other t(54)=4.009,p=0.0002.Therefore, we reject the Null hypothesis 6, and conclude that women will seek men who are more family oriented than themselves.

Discussion
Online dating has a number of important advantages over conventional offline dating.For example, it offers unprecedented (and remarkably convenient) levels of access to potential partners, which increases what Kerckhoff (1964) named "Field of Elegible".This is especially helpful for single people who might otherwise lack such access.It also allows online daters to use Computer Mediated Communications to garner an initial sense of their compatibility with potential partners before deciding whether to contact them or meet them face-to-face.In addition, refined search parameters in online dating websites or apps allow users to eliminate people who are likely to be poor relationship partners in general from the dating pool.
Online users interacting with potential partners seek to form their own impressions of their communication partners.Therefore, they must assess two things: 1.)The content of the identity claims made by potential partners 2.) the veracity of these claims.In other words, the receiver must both understand and believe the claims (Lampe et al., 2007).Deception in offline environments is common due to the ability to selectively self-present online (Walther, 1996).Therefore, some kinds of misrepresentation are easy to accomplish through CMC.
Users when attempting to meet potential partners online, rely on a variety of cues to gather information about each other.However, not all these cues are equally credible.When using online dating platforms, users do not share a social network and therefore have less access to mutual friends who could confirm or deny information about other users (Lampe et al., 2007).Some scholars researching online dating found that men tend to emphasize physical attractiveness whilst women focused on social attractiveness traits, such as kindness and intelligence.
Men and women make mating decisions very differently.Rozin & Royzman (2001) found that for men if a potential female mate is not attractive enough, she usually would not qualify for a first date.Alternatively, women are more concerned about whether or not the person is "good."Therefore, women might accept a less-attractive mate if he were outstanding in some other way.Furthermore, many women pay heavy attention on men's income.Eastwick and Finkel (2008) found that although men and women tend to say they prioritize different things in their mates, there was no difference in the types of mates people of the two sexes actually chose in a speed dating experiment.In addition, there was little association between the traits participants said they wanted in a desired partner on paper and what they actually liked about the mates at the speed dating event.This is in part due to the fact that the way in which people pair with each other on dating sites differs from the way they will later evaluate the relationship (Eastwick & Finkel, 2008).People browse online profiles in what is known as "joint evaluation mode," where they compare multiple potential partners against each other based on attractiveness, income, and other factors.However, people make relationship decisions in what is called "separate evaluation mode," by judging just the potential mate and his or her thinking.The joint evaluation model is likely to cause users to focus on certain qualities they think are important in a potential partner, perhaps to the neglect of qualities that actually are important (Finkel et al. 2012).The implications are that whilst most people may have in mind the traits they want in a potential partner, they may ultimately accept a mate who is totally lacking in these desired traits.In other words, most people tend to look at aspects that seem to satisfy their expected traits but rarely have a satisfactory method to focus on variables which can achieve a long-term relationship.
The results of the present study show certain patterns of behaviour in young Taiwanese women in terms of "background homophily."These trends include being easy going, being a risk taker, having self-control, having life plans, valuing money, and being family oriented.Our results do not support Baker's hypothesis of "like is attracted to like" (Baker, 1973;1974).For "personality" and other related variables such as "risk taking" and "self-control", most women preferred finding someone who was better than themselves.They only exhibited homophily in their value of money.These findings contradict expectations.For example, we would expect "money value", "life plan" and "family orientation" to be similar between both female and desired match and exhibit homophily, especially for older respondents.
Our study was limited by several factors, such as the sample size, the age group, and the cultural environment.Several variables were not taken into account during the present study, such as "respect", "mutual trust", "jealousy" or "control to the couple", which seem relevant in the Asian context where males and females control each other.It would be interesting to include such variables since they could be important predictors for deciding whether or not relationships are long-term or short-term.Moreover, it is interesting to evaluate men and women of the same age groups to assess how much homophily they exhibit and see potential differences according to age.It would be also interesting to test educational similarity, which may have an effect on this homophily tendency of women, since it's associated with common lifestyles and similar cultural interests which facilitate relationship development (DiMaggio & Mohr, 1985;Kalmijn, 1994;Kalmijn & Bernasco, 2001).In addition, it is possible that both "joint evaluation mode" and "separate evaluation mode" (Finkel et al. 2012) may be influenced by homophily, but it is unclear to what extent that may be true.More research is needed to better understand people's selection mechanisms and how it is influenced by homophily during online contact and face-to-face contact.The presence or absence of homophily along cultural aspects of standard beauty may also play a role on expected attractiveness before users browse profiles.Hence it is useful to further research in that direction.
Our results have business implications for online platforms in Asia or platforms with a large database of Asian users.Homophily in search results for relevant attributes can prove beneficial for their users as well as attracting Asian online daters to their services.

Conclusions
The present research assessed background homophily in young female Taiwanese college students.We found that women exhibited no tendency for homophily towards their desired partners except in the "value of money" category.There were correlations between own and desired attributes.Women sought potential romantic mates who are easier going, more risk-taking, and who have higher self-control than themselves.In addition, women preferred males who wish to have a better economic future and who want children.These findings may vary depending on the age group and cultural environment.


Importance of age in a partner

Table 2 .
Number of observations for the personalities of females and their desired matches

Table 3 .
Pearson Correlation between the personalities of the females and their desired matches

Table 4 .
Number of observations for the risk taking of females and their desired matches **. Correlation is significant at the 0.01 level (2-tailed).

Table 6 .
Number of observations for the self-control of females and their desired matches

Table 7 .
Pearson Correlation between the self-control of females and their desired matches **. Correlation is significant at the 0.01 level (2-tailed).

Table 8 .
Number of observations for the life plan of females and their desired matches

Table 9 .
Pearson Correlation between the life plan of females and their desired matches

Table 10 .
Number of observations for importance of money for females and their desired matches

Table 11 .
Pearson Correlation between importance of money for females and their desired matches

Table 12 .
Number of observations for importance of family orientation of females and their desired matches

Table 13 .
Paired Pearson Correlation between family orientation of females and their desired matches