Plenty Of Fish Experiment: Study 2

Introduction

For studing reciprocation, some dating sites such as OkCupid (see OkTrends Blog) or AYI (see Businessinsider) offer us reply rates at the first message, that is, percent the users that respond to an initial message. Furtherore when most academic papers are analyzing what characteristics are associated with contacting and being contacted by others or with replying to another’s initial message, they work generally with contacts replied at the first message too. So one tends to make the mistake of confusing response rates as first message as mate acceptance rates, which it’s wrong. Mate acceptance is not equal to rates of initial contact in which one sent a message to the other and got just a response.

Obviously most recipients refuse to send a response to those senders who don’t meet their physical requirements, so they ignore those they don’t find worthy. Say that you get an e-mail from someone, and you can tell immediately that you have no interest in communicating with that person.

So most people don’t reply at all, ever. Just delete the message. In Internet-speak, this tactic is completely understood to mean “Not interested at all, ever.”

But other users send a first short response to undesirable sender due to a protocol in terms of courtesy.

Others send a short reply saying, “Thanks for writing, but I’m not interested.” Then they delete the person’s message. If the person continues to write, they don’t answer ( if the person persists, use the blocking feature on their message system for example).

By the other hand, some people may even extend the messaging exchange with an unattractive prospect, mainly users who usually receive few or no messages from their potentially preferred partners.

But when an inicial user contacted realizes that the prospect just isn’t a match, the frequent action to stop of keeping sending more messages. Other action is to send a last message to say: “I need to stop now. I’ve enjoyed chatting with you, but I don’t think we’re a match. I don’t want to waste any more of your time. Best of luck in your search.”

So how can we distinguish real mate acceptance of reply rates at first message? or even reply rates at several messages?. Hitsch et al. (2010) explain a reliable way to measure mate acceptance rates, or as they define, matches:

[…Since we can track only the users’ online behavior, we do not know whether two partners ever met offline or eventually got married. However, our data do allow us to observe whether users exchange a phone number or e-mail address, or whether an e-mail contains certain keywords or phrases such as “get together” or “let’s meet.” We therefore have some indirect information on whether the online meeting resulted in an initial match, i.e., a date between the users. We therefore define a match as a situation where both mates exchange such contact information (i.e., for a match it is not enough for a man to offer his phone number; we also require that the woman respond by sending her contact information)…]

In other words, we can define as match or mutual acceptance when users are sharing personal information (emails, phone numbers, social networking, Skype ID, etc) with the purpose of set up an date. They further argues that:

[…For some users we observe multiple matches. We use each of these as a separate event when we describe the sorting patterns below. Matches are very different events from first-contact decisions: Table 4 shows that of all first contact e-mails sent by men, 4.3 percent lead to an eventual match…].

In other words,  match or mutual acceptance is definied if two users are sharing personal information (emails, phone numbers, social networking, Skype ID, etc) with the purpose of set up an date. They also observed a statistical pattern for all exchanges that could be defined as a match. The mean number of exchanges required until the relationship resulted in an offline date were six messages:

[…Typically, a match is achieved only after multiple rounds of e-mail exchanges: the median and mean numbers of e-mails sent back and forth before a match is realized for conversation initiated by men are 6 and 11.6, respectively…]

d

Source: “Matching and Sorting in Online Dating” ( Hitsch et al, 2010).

Thus, Hirsch et al. are providing an interesting correlation to studying those databases where researchers can not access into the contents of the messages exchanged. This makes it possible to obtain reliable measures of acceptance rates in online dating experiment. Kreager et al (2014) allude to these correlations in their own online dating study:

 

[…As the number of reciprocated responses increased, the percentage of messages in each category declined, so that only 3% of men’s, and 7% of women’s, sent messages resulted in more than five exchanges. The last category captured the mean number of exchanges (six) required until the relationship resulted in an offline date by Hitsch et al. (2010a). Obviously, the likelihood of any given message resulting in a reciprocated exchange and eventual date is extremely small..].

 

Method:

In the second study I’m going to measure response rates at the second message. It would be too laborious perform a message exchange for each pairwise to determine whether each initial recipient is willing to provide to each sender her phone number and set a date. I assume this estimate is quite interesting, and would give us a response rates data
fairly close to real rates of mate acceptance.

The dummy profiles will be introduced on the same mating system, the dating site “Plenty of Fish”, over the same 1-week period. They were placed into the other geographical place, a big-sized city named B.

MESSAGING STATISTIC:

Messages received: Number of messages received per time active on site. (7-days).

Contacts initiated:  Number of indivuduals I’ve contacted through each dummy profile.

These messaged users are not randomly chosen. I chose a sample of 100 female users for each dummy through advanced search. They were to meet the following parameters:

Age range: 18-35 years old.

City or Postal Code / Miles: 100 miles

Body type: “thin”, “athletic” and “average”. I discarded profiles self-identified as “a few extra pounds, “big & tall / bbw” and “prefer not to say”.

Within the subset of browsed female profiles, I chose as target those profiles that contained pictures from women that could be included within the upper half of facial attractiveness spectrum. This is, since avarage or medium-attractiveness until highly-attractiveness ( 5 to 10 oints on a 1-10 scale).

Contacts received: Number of people who contacted this person per time active on site.

Reply proportion at first message: Proportion of initial contacts from this person to which others replied one.

Reply proportion at second message: Proportion of initial contacts from this person to which others replied twice.

Localitation: big-sized city B.

Data-collection period: 7- days.

Dummy profiles:

This second study is composed of 4 male profiles, 2 of them highly attractive (males Z and Y; r(>8), 2 moderately attractive: male F (r= 6.93), male C (r=6.34).

Recall that to ease attractiveness comparisons, I’ve sorted the ratings into five equal categories of attractiveness (highly-attractive: 8-10, moderately-attractive:6-8, medium-attractive:4-6, and below medium-attractive(< 4).

I assigned higher occupational and educational levels to less physically attractive profiles. The highly attractive profiles were assigned with blue-collar occupations, while the rest of the profiles were configured with white-collar professions (see study 1).

Male Z:  Fitness&Kickboxing trainer; Male Y: Nightclub Promoter; Male F: Architect; Male C: Mathematical.

a) High-attractiveness dummy profiles (>8 points on a 1-10 scale):

  • Male Z :
Highly attractive male Z

Figure 1. Photos of male Z

Male Z Profile

Figure 2. Headboard of Male Z Profile

  • Male Y:
Highly attractive male Y

Figure 3. Photos of male Y

 

Male Y Profile

Figure 4. Headboard of Male Y Profile

 

b) Moderately-high attractiveness dummy profile:

  • Male F
Male partner F

Figure 5. Photos of male F

Male F Profile

Figure 6. Headboard of Male F Profile

  • Male C
Male partner C

Figure 7. Photos of male C

perfil

Figure 8. Headboard of Male C profile.

 

Results:

a) High-attractiveness dummy profiles (>8 points on a 1-10 scale):

  • Male Z :

1) First screenshot (third day):

100 enviados al comienzo

Figure 9. Screenshot showing Contact History: Contacts initiated.

22 contactos iniciales

Figure 10. Screenshot showing Contact History: Contacts received.

Mail inbox = 65

Meet me = 23

Contacts iniciated = 100

Contacts received = 22

Reply proportion at first message =  65-22 = 43 out of 150.

Reply rate at first message = 43 %

2) Second screenshot (sixth day):

final mensajes enviados

Figure 11. Screenshot showing Contact History: Contacts initiated. The number of females contacted were 100. Note that the screenshot after the 7 day period shows 99 users, because 1 of them has removed her account.

final contactos d

Figure 12. Screenshot showing Contact History :Contacts Received.

Mail inbox = 62 (9 of them were first responses. So 9 female users responded to the first contact but after the first screenshot). I won’t send a second message to these delayed 9 users, since the possible answers will be out of data collection period of 7 days.

Meet me = 36

Total contacts received = 32

New contact received = 10

Reply proportion at first message revised =  43 +9 = 52 out of 100.

Reply rate at first message = 52 %

Reply proportion at second message = 43 out of 100. (*) The data is unreliable because 9 users who responded to the first message did too late and I could not record their behavior for a second message.

Reply rate = 43% (*)

3) Contats received and Meet me at seventh day:

final contactos 42

Figure 13. Screenshot displaying Contact History: Final Contacts Received

Meet me at 7 day.

Figure 14. Screenshot disyplaing Meet me at 7 day.

Meet me = 56.

Total contacts received = 42.

 

Male Y :

1) First screenshot (third day):

20 contactos

Figure 15. Screenshot showing Contact History: Contacts Received:

69 y 16 meet

Figure 16. Screenshot showing Contact History: Contacts Inicitated. The number of Contacts Initiated was 100 (I made a wrong screen capture).

Mail inbox = 69

Meet me = 16

Contacts initiated = 100

Contacts received = 20

Reply proportion at first message =49 out of 100.

Reply rate at first message = 49 %

2) Second screenshot (sixth day):

98 mensajes enviados finales

Figure 17. Screenshot showing Contacts Initiated. The number of females contacted were 100. Note that the screenshot after the 6 day period shows 98 users, because 2 of them have removed their accounts.

23 contactos final

Figure 18. Screenshot displaying Contact History: Contacts Received.

Mail inbox = 45

Meet me = 28

Total contacts received = 23

New contact received = 3

Reply proportion at second message = 45-3 = 42 out of 100.

Reply rate at second message = 42 %

3) Contats received and Meet me at seventh day:

f2

Figure 19. Screenshot displaying Contact History: Contacts Received at 7 day.

f

Figure 20. Screenshot displaying “Users who want to meet you”.

Meet me = 45

Contact received = 45

b) Moderately-high attractiveness dummy profile:

  • Male F

1) First screenshot (third day):

18 mensajes

Figure 21. Screenshot showing Contact History: Contacts Inicitated.

contactos iniciales

Figure 22. Contact History: Contacts Received.

Mail inbox = 18

Meet me = 8

Contacts initiated = 100

Contacts received = 3

Reply proportion at first message =15 out of 100.

Reply rate at first message = 15 %

2) Second screenshot (sixth day):

100 mensajes finales

Figure 23. Screenshot showing Contact History: Contacts Initiated. The number of females contacted were 100. Note that the screenshot after the 6 day period shows 99 users, because 1 of them have removed her account.

Contacts recieved

Figure 24. Contact History: Contacts Recieved

Mail inbox = 5

Meet me = 10

Total contacts received = 4

New contact received = 1

Reply proportion at second message = 5-1 = 4 out of 100.

Reply rate at second message = 4 %

3) Contats received and Meet me at seventh day:

There were no changes,

Meet me=10

Total contacts received = 4

 

  • Male C

1) First screenshot (third day):

1b

Figure 25. Contact History: Contacts Received

1

Figure 26. Contact History: Contacts Inicitated

Mail inbox = 14

Meet me = 2

Contacts initiated = 100

Contacts received = 3

Reply proportion at first message =11 out of 100.

Reply rate at first message = 11 %

2) Second screenshot (sixth day):

2 a

Figure 27. Contact History: Contacts Inicitated

2b

Figure 28. Screenshot showing Contact History: Contacts Received. Note that the screenshot after the 6 day period displays 1 users, because 2 of the inicital contacts have removed their accounts.

 

Mail inbox = 1

Meet me = 3

Total contacts received = 3

New contact received = 0

Reply proportion at second message =1 out of 100.

Reply rate at second message = 1 %

3) Contats received and Meet me at seventh day:


f

Figure 29. Screenshot showing final number of Likes in “Meet Me” at 7 day.

Meet me = 4

There were no changes in contacts received.

 

Table of results:

 

dd

(*)= Reply proportion at second message = 43 out of 100. This data is unreliable because 9 users who responded to the first message did too late and I could not record their behavior for a second message.

 

Dating pool in meet me:

The highly-attractive men has as mean 50.5 women interested (potential matches in “meet me” section) per week, 7 times more women interested than the moderately-attractive males ( mean: 7 women).

Contacts received:

The four accounts combined received 97 unsolicited messages, and two best looking men monopolized 99.6 % of these contacts.

Reply rate at second message:

The mean response rate for the two most attractive men is of 45.5%, quite higher than the response rate for moderately-attractive profile (2.5%). Or in other words, best attractive guys are reciprocated by almost than 1 out of 2 women. While the moderately-attractive males are reciprocated by around 1 out of 40 women.

In the third study I will complete the database, mainly experiencing with several profiles within the medium-attractiveness spectrum and one below-average attractiveness man.

 

 

 

 

 

 

 

 

 

 

Advertisements
This entry was posted in mate choice, Mutual Match, online dating, pof experiment, reply rates, Uncategorized. Bookmark the permalink.

10 Responses to Plenty Of Fish Experiment: Study 2

  1. thegreatshebang says:

    That is a lot of work, but as usual the results are fascinating.

    • sirtyrionlannister says:

      Thank you. It is indeed a lot of work invested, but hopefully worthwhile somehow.

      • thegreatshebang says:

        Regarding primary research, it’s never easy to predict the impact of the results, especially when you can’t predict (although you can hypothesize) the results ahead of time!

        One interpretation of this study is that men “should” not make much effort to date. It is considered normal, or normative, for men to make a lot of effort to find a mate. But the men in the high attractiveness category don’t need to make much effort, comparitively, and men in the less attractive categories hardly get any results at all!

        It struck me too that men in the lower categories wouldn’t know if the woman that did return their message was also trying to communicate or meet with several high attractive men. He might be in a dating pool, in the eyes of one woman, of 6 men where he is on the edge of her threshhold. I think that would be hard to answer unless you got the funds to interview the women directly.

        I do think your conclusion that for many men the dating process is like a lottery is based on more and more of your evidence.

  2. sirtyrionlannister says:

    While choosy women (higher rate thresholds) are selecting within a narrow specific subset of potential males, less-choosy females (e.g. those replying back some moderately-attractive and medium-attractive male) are picking out within a wider pool of potential mates. This is because the number of potential mates that are acceptable will increase as the acceptance threshold is decreased. Which in turn is reducing the options to get a real match with a permissive woman (without also recording phenotypic variation among receptive and discriminating women in dating landscape), because of the high degree of competence in their socio-sexual niche.

    More than a lottery could be said that there is a high degree of determinism in mate choice. So it seems inappropriate to invoke a probabilistic theories, and I am in favor of discarding old colloquial fallacies like “dating is a numbers game”, etc…

    If sometimes mating systems are slightly stochastic and non-entirely deterministic is due to some external events that we can hardly control. The mate preference functions for phenotypic traits are usually kept constant in a same socio-ecological environment.

  3. wub1234 says:

    Firstly, I commend you for this extremely well researched, written and argued site, the overwhelming majority of which I agree with. I just wanted to bring a few points to your attention where I perhaps slightly disagree, or where I think you might have overlooked something. I’ll try to do this as succinctly as possible.

    In the debunking of women going for older men, I think it is notable that women hardly ever go for anyone younger, and particularly much younger. Wouldn’t you logically expect to see 35 year-old women going for good looking 25 year-olds (who would then reciprocate due to the diminishing dating options)? This doesn’t happen, or only in very rare cases. Women are five times more likely to pair off with someone ten years older than ten years younger, and actually the majority of women pair off with someone older.

    In terms of the majority of marriages being between people of the same age, I agree that this is statistically significant, but I would suggest an alternative explanation for this. Many couples meet at school, college or university where we are grouped together based on age, so this distorts the figures.

    When dealing with online dating, it is notable that we get very little to go on other than a picture. When I’ve done online dating, I also completely judge women by this picture! I have higher standards than many men, and so am more selective. I have still managed to secure numerous dates though online dating, although I do have the advantage of being a professional writer! But all of the usual social cues that we rely on to judge people don’t exist online, so naturally we just go for the best looking people, unless someone manages to write something particularly engaging.

    Additionally, male strategy with online dating (send out as many messages as possible) dictates that women are bombarded with messages. So they have to adopt some sort of strategy to sift through them, and as most of them are bollocks, they naturally pick the best looking guys. This would happen anyway, but the approach of men, the extent to which they will hit on anyone both online and in real life, and the general lowering of standards, all contribute to this.

    I also think you have underestimated the importance of socio-economic status. I can only provide anecdotal evidence for this, but…I happen to live in one of the wealthiest parts of the UK. I can tell you now, it is an absolute MILF goldmine here! Where I live, you hardly ever see a woman who is overweight, I’ve barely seen one since I lived here. Certainly looks play a major role, but you shouldn’t understate the importance of socio-economic status. Just for starters, many professional women will rule out dating non-professional men (and there are both dating sites and introductory agencies catering for this).

    It’s harder for me to judge the looks of men, but I would suggest that where I live men have been more able to trade upwards in terms of looks, or achieve an approximately equivalent mate, than would be the case in a poorer suburb, or just generally in society. And I’m quite sure this will be reflected in any wealthy neighbourhood.

    So my estimation of female attraction is that it’s 90-95% based on looks and socio-economic status, and that the proportions alter somewhat over time generally. When women are younger, looks and popularity are absolutely key. As they get older, social status and money become more important. Probably never more important than looks, but if you walk around the streets where I live you might come to question that assumption.

    Nonetheless, I commend you on your site. Please keep up the good work.

    • sirtyrionlannister says:

      @ wub1234

      Thanks for your comment,

      “In the debunking of women going for older men, I think it is notable that women hardly ever go for anyone younger, and particularly much younger. Wouldn’t you logically expect to see 35 year-old women going for good looking 25 year-olds (who would then reciprocate due to the diminishing dating options)? This doesn’t happen, or only in very rare cases.”

      Yes it happens, and that propensity is easily observable today (which could be labeled colloquially as “cougar trend”) whenever one check out female dating profiles and their stated age preferences and messaging stats. I mean women over or close 30-35, and this kind of trend is increasing more sharply in older age groups.

      “Women are five times more likely to pair off with someone ten years older than ten years younger, and actually the majority of women pair off with someone older”

      It seems you’re confusing mating preferences (Sampled response to stimuli) with mating pattern (Some observed stats in marriages and cohabiting couples. Which is moreover ruling out short term relationships and occasional mates).

      “In terms of the majority of marriages being between people of the same age, I agree that this is statistically significant, but I would suggest an alternative explanation for this. Many couples meet at school, college or university where we are grouped together based on age, so this distorts the figures.”

      Yes at school or college, when these young girls are exactly grouped into that age range where they are still feeling attracted to men his age or a few years older. For example you could try to look at divorced women in their 30s or 40s or those that are still single and then try to find out which are their age preferences in potential mates.

      “When I’ve done online dating, I also completely judge women by this picture! I have higher standards than many men, and so am more selective. I have still managed to secure numerous dates though online dating, although I do have the advantage of being a professional writer! But all of the usual social cues that we rely on to judge people don’t exist online, so naturally we just go for the best looking people, unless someone manages to write something particularly engaging.”

      You seem to describe online dating as a parallel universe. I rather think that is a fairly accurate reflection of actual mate preferences and sampling tactics (decision rules adopted). True that perhaps other ecological landscapes lead to a higher assortative mating due to social constraints, greater self-awareness of competition over number of mates, choosiness (investment in mate choice), etc.

      “Additionally, male strategy with online dating (send out as many messages as possible) dictates that women are bombarded with messages. So they have to adopt some sort of strategy to sift through them, and as most of them are bollocks, they naturally pick the best looking guys. This would happen anyway, but the approach of men, the extent to which they will hit on anyone both online and in real life, and the general lowering of standards, all contribute to this.”

      I agree with these statements. A surplus of male suitors (overabundance of unsolicited messages) for each women produces a increase of ‘choosiness’, the effort an female is prepared to invest in mate assessment. And matterns of mate choices can be altered by changing the costs of choosiness without altering the preference function.

      “I also think you have underestimated the importance of socio-economic status. I can only provide anecdotal evidence for this, but…I happen to live in one of the wealthiest parts of the UK. I can tell you now, it is an absolute MILF goldmine here! Where I live, you hardly ever see a woman who is overweight, I’ve barely seen one since I lived here. Certainly looks play a major role, but you shouldn’t understate the importance of socio-economic status. Just for starters, many professional women will rule out dating non-professional men (and there are both dating sites and introductory agencies catering for this).”

      Being a high status male (with respect to mating), nowadays, says less about material wealth or social status, than about physical beauty. From my empirical experience, I honestly believe that the role of social status on mating access is marginal in overall terms. But there is clearly a small evolutionary niche where a certain amount of women (i.e. gold diggers) have always been prone to mating with men of high status (i.e. celebrities) , regardless of their physical appearance.

      “So my estimation of female attraction is that it’s 90-95% based on looks and socio-economic status, and that the proportions alter somewhat over time generally. When women are younger, looks and popularity are absolutely key. As they get older, social status and money become more important. Probably never more important than looks, but if you walk around the streets where I live you might come to question that assumption.”

      I will not argue against, I do not doubt that assumption. All things being equal, women will favor wealthy/high status males, but only in very exceptional cases (often involving very high profile individuals in the public eye – like those uglies male celebrities being able to enjoy beautiful ladies-, tending to skew perceptions of normal) will a female make significant concessions in terms of the physical attractiveness of her mate (unless she is not, herself, attractive enough to warrant the attentions of physically attractive males).

      Not so surprising, when one considers that a selection bias for resourceful males should exist in some proportion to the advantages they pose to the survival of her offspring. Thus, if the advantages are small (given a prosperous welfare state, which marginalizes these advantages), then there will be minimal selection bias (which explains a large population of women who are increasingly disinterested by the lone prospect of a resourceful mate).

  4. Travis says:

    Allthough the medium attractive men didn’t get as many messages as the highly attractive males, they definitely have options and won’t struggle in the dating game. How attractive were the women that messaged the medium attractive men?

    • Sir Tyrion Lannister says:

      @ Travis,

      Recall that the profiles corresponding to those men rated as medium attractiveness are: A, B, D and E.

      Their response rates, received unsolicited messages and female “likes” computed among the three studies were included in the following table:

      In some screenshots made for the section “They made first contact” are shown female thumbnails, take a look. Anyway as far as I can remember there was no attractive woman (i.e. when I whatched their enlarged photos). Neither among those women who said click “yes” on “meet me”.

      • Travis says:

        @Tyrion

        Yeah I noticed that at the last second before I sent that comment. They’d probably be better off looking for a woman at college,social circle, recreational activities etc. Definitely would be the top 1-3 guys in most circumstances.

        • Sir Tyrion Lannister says:

          @ Travis,

          Although women feel a profound neurological bias towards the most conspicuous male phenotypes, sometimes they make assesment errors. Moreover we don’t know the sampling done by each indivual to arrive at a mate-choice decision and thus the outcome of any female decision could be stochastic.

What do you think? Let me know in the comments below!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s