Eine interessante Studie zu „Intelligenz vs Attraktivität“ auf dem Dating Markt:
We study partner preferences for education and attractiveness by conducting a field experiment in a large online dating market. Fictitious profiles with manipulated levels of education and photo attractiveness send random invitations for a serious relationship to real online daters. We find that men and women prefer attractive over unattractive profiles, regardless of own attractiveness. We also find that higheducated men prefer low-educated over high-educated profiles as much as high-educated women prefer high-educated over low-educated profiles. With preferences similar for attractiveness but opposite for education, two groups are more likely to stay single: unattractive, low-educated men and unattractive, high-educated women.
Quelle: Brains or beauty? Causal evidence on the returns to education
and attractiveness in the online dating market
Aus der Studie zu den Schwierigkeiten:
It has proven difficult to credibly identify partner preferences. One difficulty arises because of search frictions. Individuals generally self-select into clubs, schools, jobs, and neighborhoods. If search frictions make individuals more likely to meet their partner in self-selected environments (Kalmijn, 1998), they are also more likely to end up having a partner with traits similar to their own, regardless of partner preferences. Another difficulty arises because of psychological and social frictions. Partner choices are based on a combination of rationality and emotions (Finkel et al., 2012). If fear of rejection leads disappointment averse individuals to act strategically, they may shy away from partners they consider attractive but unattainable (Bell, 1985; Loomes and Sugden, 1986; Gul, 1991). In fact, psychological research has documented that (some) people are indeed rejection sensitive, which ultimately affects their search for partners and intimate relationships (Downey and Feldman, 1996; Downey et al., 1998). Importantly, because of the fear of rejection, realized partner matches will not necessarily reflect partner preferences, even in the absence of search frictions.
A final difficulty arises because of correlated traits. Any observable trait similarity between partners could be due to similarities in other related traits that are observable to potential partners but not to researchers. Whenever observable and unobservable traits are correlated, either between or within partners, the estimated partner preferences for specific traits will be biased.+
In der Tat kann es sehr schwierig sein heraus zu finden, was genau der Grund war, weswegen sich jemand für einen Partner entschieden hat. Vielleicht hat sich der Partner auch für ihn entschieden und er war froh überhaupt jemanden gefunden zu haben. Oder es liegt eine Eigenschaft vor, die eben andere interessant finden, derjenige aber nicht. Und natürlich sind auch künstliche Experimente mit fiktiven Entscheidungen schwierig, weil sie teilweise nur schwer etwas wie Ablehnung oder positive Zeichen oder „Kosten des Gesamtpakets und wo macht man welche Abstriche“ richtig simulieren können.
Klar will jeder einen hübschen, intelligenten, stark in einen verliebten, fürsorglichen, mit einem im sexuellen Bereich kompatiblen Partner, der einen vernünftigen Job und einen guten Charakter hat, aber wie weit reduziert man das in welchem Bereich, was ist „noch gut genug“ und was wiegt wieviel.
Zum Ansatz hier:
Our aim is to construct twelve fictitious profiles that appear as similar as the real ones. Our fictitious profiles are born in 1985, live in Amsterdam, are heterosexual, not religious, of average weight and height, live an active life (no smoking and regular exercise), and aim at a long-term relationship with children.8 Our fictitious profiles do not provide occupation and income information. There are two points to note about our profiles. First, profiles with an explicit demand for a long-term relationship and children in combination with the birth year (people in their early 30 s) signal the desire for a serious relationship. Second, profiles without occupation and income information are realistic profiles: that is, the online profiles under study do not offer income options, and most profiles leave the occupation option open. Also, profiles without occupation and income information allow us to interpret the effect of education, for example, more broadly as the effect of education and everything else that education signals, including occupation and income status. Almås et al. (2017) refer to education as one potential measure of someone’s earnings potential and we use the same terminology here.9
We use six profiles for each gender (see Table 1). The six profiles have different usernames, education levels and photos. We assign different usernames based on popularity ratings of Dutch names for the 1985 cohort.10 The profiles only hold highly ranked usernames to overcome the concern that online daters may respond differently to first names that appear odd and/or reflect social background (Fryer and Levitt, 2004). We assign two levels of education. The low-educated profiles hold a (vocational) high-school degree and the high-educated profiles hold a university degree. We choose these two categories to create a significant distance in terms of education, while at the same time making sure that the educational groups are not too marginal (e.g., high school dropouts or people with doctor degrees).11 We finally assign six profile photos (for each gender) representing three different attractiveness levels. For each education category we use one profile photo capturing low attractiveness, one photo capturing medium attractiveness, and one photo capturing high attractiveness.
Da kann es also nur erst einmal um ein erstes Interesse gehen. Es wurde dann wie folgt vorgegangen:
Next, we hired a professional data scraper, who built a robot that browsed the website, identified the relevant sample, and downloaded each subject’s profile photo, education level, area of residence, age and gender. The photos and education information are crucial for analyzing matching preferences in the attractiveness and the education dimensions. The scraping produced a sample of 2,672 (1,554 male and 1,118 female) online daters. Because five profiles were incompletely scraped (in particular, they did not have a profile picture), we work with a sample of 2,667 (1,550 male and 1,117 female) online daters with complete records who are eligible for receiving an invitation message from one of our 12 fictitious profiles.
und dann weiter:
The online dating members in our sample were randomly assigned to one of the six profiles of each gender from whom they receive a message.19 The profiles sent out identical messages, expressing their interest in the subject and clearly stating that they are looking for a long-term relationship. The message also includes an invitation to meet in real life. The invitation message is found in Appendix C. Each profile sent out roughly 120 messages in each round. For technical and practical purposes (i.e., passing the initial dating site screening), we restrict the messages sent from each profile to blocks of 20 per day, using various computers and locations for different profiles. We further post each profile with a one-day lag.
Also sie haben dann die Leute angeschrieben unter Verwendung der Profile. Dann bewerten sie die Antworten als Positiv oder negativ, quasi als Indikator für Interesse oder kein Interesse:
Table 4 summarizes the data on response behavior that we collected within the first four weeks after the initial invitation message. About half of the invitations were opened. The response is recorded as being positive as long as it is not a definite rejection.
The shares responding and responding positively are almost similar, showing that online daters basically only respond when they are interested. We use a positive response as the main outcome variable throughout. In line with earlier research (e.g., Hitsch et al., 2010), we find lower response rates for women (13 percent) than for men (29 percent). Women are also slower when responding: among the responders the average wait time is 1.90 days for men and 2.99 days for women (although both the median man and woman responds the same day the letter is sent). Online daters who respond positively receive a polite rejection message from the profile indicating that they are no longer interested. The rejection message is found in Appendix C.
Aus den Daten:
Wie man links sieht antworten Männer bereits häufiger und eher bei High School als bei Universität.
Bei den Männerprofilen unterscheiden die Frauen nicht viel zwischen High School und Universität
Körperliche Attraktivität hat bei beiden den zu erwartenden Stufenverlauf, wobei Männer deutlich großzügiger sind.
Aus der Studie:
Fig. 1A demonstrates the importance of attractiveness in the dating market. We see that high-attractive profiles receive many more responses than intermediate-attractive profiles, and that intermediate-attractive profiles receive many more responses than low-attractive profiles. In line with the existing literature, women are much less likely to respond to profile invitations than men.20 For example, 20 percent of the men respond to the invitation from the low-attractive profiles whereas less than 4 percent of the women do so. However, the relationship between response rates and attractiveness appears to be equally steep for male and female online daters.
Turning to education, Fig. 1B shows interesting gender differences. Women (on average) do not care about the level of education of the profile partner: their response rates are almost identical for the two male profiles. In contrast, men seem to care, and persistently prefer low-educated profiles over high-educated profiles.
Table 5 contains the estimates from our response regressions. In columns 1 and 2, the estimates echo what we already saw in Figs. 1A and 1B: response rates of both men and women significantly rise with profile attractiveness, and response rates significantly fall with profile education for men, but not for women.
For profile attractiveness, we find that profiles with high attractive photos are almost 20 percentage points more likely to receive a response than profiles with low-attractive photos, regardless of gender. These are large effects: a 20 percentage point rise in response rates represents a 200 and 600 percent rise in male and female responses, respectively. These latter differences arise because of the large gender differences in the baseline response rates.
In column 3 we test for gender differences in response rates measured at the margin. While we find that male daters are somewhat more responsive to intermediate-attractive profiles than female daters, we believe that the high p-values for the high attractive profiles indicate that attractiveness is more or less of equal importance to men and women.
For profile education, we find that female daters do not respond to profile education: the point estimate is close to 0 and statistically insignificant. In contrast, we find that men are 5.1 percentage points less likely to respond when we increase the education level of the female profile from high-school to university, which is equivalent to 25 percent fewer responses.21
In the remaining columns, we also test for interactions between attractiveness and education. We find that none of the interactions is statistically significant, neither for male nor for female daters.
Apparently, online male daters are not more forgiving towards female profiles with high levels of education when these females are more attractive. Education and attractiveness seem not to be substitutable. Likewise, we find no evidence of complementarities between attractiveness and education either.22
Das ist das Problem aller Studien aus dem Bereich. Sie werten nicht tatsächliches Interesse nach einem Kennenlernen aus, sondern erstes Interesse. Dabei werden Frauen erst einmal darauf schauen, ob sie denjenigen attraktiv und sympathisch finden, alles andere kann man dann im Gespräch sehen.
Der Aufbau über erstes Interesse ist gut für eine Studie, weil es die Vorgänge vergleichbar macht. Würde man tatsächliches Dating analysieren, dann wäre das zwar interessanter, man müsste aber die gesamten Gespräche analysieren, was sehr subjektiv ist.
Interessant ist, dass Männer höhere Ausbildungen eher aussortieren.
Dazu aus der Studie:
Here we present a collage of recent empirical evidence consistent with, and complementary to, these preference-based predictions. The traditional partner matches we predict are mutually preferred by both partners and rather consistent with the traditional partner matches persistently observed in marriage markets with more educated women than men (Bertrand et al., 2015; Folke and Rickne, 2020, and Almås et al., 2017).
While we might be surprised that high-educated men preferlow-educated women, which undeniably lowers the expected household income, traditional preferences predict that some men are willing to forego such income gains in return for a less educated partner with a weaker attachment to the labor market. Having a less educated partner may make work-family trade-offs easier to motivate. The reverse logic may also explain why higheducated women turn down low-educated men: high-educated women that want to prioritize family over their career can better motivate that trade-off by having a partner with the same (high) education-level. Interestingly, a survey from the Netherlands Institute of Social Research confirms that traditional households are still the prevailing norm for Dutch households; of all heterosexual households were at least one partner is working full-time, the man is the only full-time employed in 76 percent of the households, whereas both partners work full-time in 15 percent of the households, while the woman is the only full-time employed in the remaining 9 percent of the households (SCP, 2008).
The partner mismatches we predict are also consistent with real world behavior of unmatched men and women. For unmatched men, there is evidence that the least educated men try to find their partner elsewhere. Glowsky (2007), for example, shows that loweducated men in more developed countries increasingly marry women from less developed countries. For unmatched women, there is also evidence that high-educated women find it difficult to find a partner and get pregnant. Recent studies on singles in the U.S. suggest that in many marriage markets there are too few college-educated men to meet the partner demand of collegeeducated women (Klinenberg, 2012; Birger, 2015), and that female college-graduates shade their career ambitions to possible partners (Bursztyn et al., 2017). Other recent studies on childlessness in women find the highest share of childless women, as well as the highest IVF (in vitro fertilization) treatment rates among collegeeducated women (Bitler and Schmidt, 2012; Jalovaara et al., 2017; Lundborg et al., 2017). Although these studies do not consider childlessness under single women, we believe that childlessness is closely tied to being single.
Im Ganzen aus meiner Sicht eine eher etwas enttäuschende Studie, die ich den Abstract nach interessanter fand.