Psychological experiments without permission of dating SNS, it works well if it is told that it is the 'best partner' even if it is not compatible


by

Candida.Performa

Facebook conducted a psychological experiment without permission from the user and officially apologized, but OkCupid , a dating SNS, told the user without permission, 'What happens if you tell two people who are incompatible with each other?' I made it clear on my blog that I was conducting an experiment, 'What happens if I don't display the profile text?'

We Experiment On Human Beings! «Ok Trends
https://www.gwern.net/docs/psychology/okcupid/weexperimentonhumanbeings.html

The experiments that OkCupid allegedly conducted are 'remove all photos from the site for a certain period of time to see what kind of changes there are', 'hide the text in the profile column of some users', 'good compatibility' Tell the user 'incompatible' and tell the incompatible user 'incompatible'. '

◆ Experiment 1: Love is blind


by

Karrie Nodalo

OkCupid was developing an app to make a blind date in the first place, but this idea that the staff thought was 'genius' failed, and the app developed over a year and a half took less than 6 months. It disappeared from the Store.

However, to celebrate the fact that we released an app with a self-proclaimed genius idea that 'you can date without knowing who the other person is', January 15, 2013 was called 'Love Is Blind Day'. I will do a certain experiment. I took all the photos from OkCupid, a dating SNS that is currently open to the public, and created a 'Blind' state.

As a result, the following becomes clear.

7 hours after the photo was removed, the response to the message increased by 44% compared to a normal day.
・ The conversation lasts for a long time
・ Contact is made at a better tempo
・ In short, OkCupid has been improved.

Below is a graph of 'Number of new conversations starting per hour' on regular days and Love Is Blind Day. On normal days, the number of new conversations peaks before 6am and tends to increase from there, but on Love Is Blind Day, new conversations slammed at 9am when the photos were removed. It decreases, then conversations with certain people continue, and conversations with new people surge at 4 pm when the photo is posted. In other words, while the photo wasn't posted, I kept talking to one person without looking for a new person.



However, this does not mean that 'look is not important'. When the photo was posted again at 4 am, the magic of the people who were talking without seeing the other person melted and the search for a new person began. 'It was as if I suddenly turned on the light at the bar at midnight,' the blog says.

In response to this result, looking back on the data when OkCupid released the blind date application, it was found that 'Once you get to the date, you are more or less satisfied regardless of the appearance of the other party.' It turns out.



Curiously, we also know that for women, the so-called 'good guy' is less satisfied after a date. According to OkCupid's hypothesis, this is because a 'good guy' is likely to be a 'fucking bastard.' We also know that these trends are common to both offline and online.

◆ Experiment 2: Is photography so important?


by Markus Spiering

Like many websites, OkCupid employs a system in which other users rate a user's profile. However, what is a little different from the others is that the evaluation is done on the two axes of 'individuality' and 'appearance'. At the right end of the image below, there are two places to evaluate 'LOOKS' and 'PERSONALITY'.



How do people evaluate people on the two axes of appearance and individuality? The result is the graph below. The horizontal axis is the appearance and the vertical axis is the individuality, but the set of points draws a graph that is almost proportional, which means that the 'appearance' and 'individuality' were the same for the user.



'Which of the profile pictures and texts influences the rating?', OkCupid erases the text in the profile section of some users, and what happens? I observed. When we experimented with two types of users, one who publishes photos and texts on their profile, and the other who publishes only photos on their profile, it was found that the percentage of profile texts affecting the rating was less than 10%.



In other words, no matter how great the words are, you may not be able to beat a single photo.

◆ Experiment 3: Power of advice


by Sean Molin

The third experiment is about the 'compatibility diagnosis' message that OkCupid gives to users. OkCupid analyzes the success of message exchanges, conversation length, information exchange, etc., predicts compatibility between users in a unique way, and conveys the results to users. Is Compatibility Diagnosis Really Working? So, when I told the two people who are incompatible with each other as 'the best partner', the users sent messages to each other. However, this isn't something to note, as OkCupid also gives users the advice, 'It's a good match, so let's exchange messages.'

However, it was said that not only did they send messages to each other, but the people who told them that they were compatible with each other became even more intimate. People can get intimate when they say 'it's a good match', even if it's not really a good match.

This result worried OkCupid members. The compatibility diagnostic algorithm developed by OkCupid is garbage, because people could just be driven by suggestion. So OkCupid then conducted an experiment that 'tell the two people who are actually good at it that they aren't good at it.'

The results are as follows. If you tell two people who are actually compatible with each other that they are compatible, the probability of starting a conversation is as high as 20%, which is a natural result. However, even if you tell two people who are compatible with each other that they are not compatible with each other, there is a high probability that they will be connected, and the significance of the compatibility diagnosis remains questionable.



Regarding the series of experiments, OkCupid explained that 'as long as you use the Internet, users are subject to hundreds of experiments at any site at any time,' and although unauthorized experiments are controversial, Facebook expresses user's feelings. While we experimented to change it, OkCupid is experimenting to improve its services, and there is a view that the two are different.

in Note, Posted by darkhorse_log