Did the invention of Tinder result in an increase of casual sexual activity?
Evidence from 1.3 million college students between 2012 and 2016. 🧵
Online dating is now one of the top ways people form relationships. About 39% of relationships began on apps as of 2019.
However, rather than a relationship formation tool we often associate apps with casual sexual activity in popular culture.
Researchers examined a very large sample (1.3M) of college students between 2012 and 2016 using the National College Health Assessment (NCHA) survey to estimate a relationship between Tinder use and sexual activity.
Greek-affiliated fraternity/sorority members saw a small increase in total and monthly sexual partners over the last month, by about 6.3% and 3.4%.
Or, about .11 and .07 standard deviations.
Greek fraternity/sorority students were also more likely to contact an STD or unplanned pregnancy following the Tinder rollout - but only by 0.2%.
This effect was concentrated in the top 25% of men: 4.3% of the increase in sexual partners (of the 6.3% total).
In other words, the top 25% of men saw an increase in sexual partners of 4.3%.
The bottom 75% saw an increase in sexual partners of 2%.
Tinder may have helped the top 25% of men twice as much!
At the same time, it also helped them remarkably little.
A 4.3% increase in sexual partners is almost nothing.
This is contrary to the belief that Tinder has had a large role in facilitating casual sex.
It is more consistent with past research on dating apps that has found people primarily do not use them for casual sex:
The effect of Tinder on sexual outcomes for women was also the same (or smaller but not statistically significant) as it was for men.
Contrary to some claims that Tinder has resulted in more sex for women, but less sex for men.
A few statistical comments beyond this point:
When we correlate changes in behavior following an event, it is sometimes described as a "quasi-causal" methodology.
The idea being that you can compare before/after an event to see if the event had an effect.
Ultimately it is still a correlation however, so be cautious in interpreting "causal" this way.
Without assignment to groups it strictly can't be known if your variable had a casual effect.
A word can be said about statistical significance and effects sizes as well:
When you have a large sample, almost any change at all will be statistically significant.
Even if the change is very small, not "clinically/practically" significant, or not meaningful.
Here we see what is effectively less than one additional yearly sexual partner on average following the Tinder rollout - so a pretty good example of "statistically significant, but not meaningful."
You should always look at effect sizes.
You could also interpret this from seeing the size of the increase in standard deviations, which correspond to the Cohen's d effect size conventions: very small at .11 or .07
This is what it would look like:
What are the take home points for this?
Maybe two big ones:
1. There is surprisingly little difference in sexual activity between the top 25% of men and the top 50%.
2. The Tinder rollout was only very weakly associated with more sex.
Even among college students, who we tend to think of as highly sexually active - and even for the Greek frat/soror students we tend to think of as the most popular and sexually active!
Why could this be - certainly the most attractive men could have more sex?
People likely underestimate the general trend (especially for handsome men) of forming monogamous relationships.
Rather than pursuing high levels of promiscuity, more common is entering a relationship.
Indeed - we see a large majority of sex occurs in relationships with a committed partner in representative data:
A common trajectory for handsome men on dating apps: they find a relationship easily, but are taken off of the sexual market as they enter into that relationship.
Thus, keeping absolute partner count lower than it might have been had they pursued a promiscuous strategy.
Here are the results of a recent poll. I asked people to imagine 100 dating-age peers and tell me if at least half were attractive enough to date.
Most people don't find most people attractive.
So - how is it that most people are in relationships? 🧵
This is a result I expected. Most people don't find most people attractive.
Women are also more selective than men.
This raises a question - how is it that so many people are in relationships and married when we don't find most people physically attractive enough for a relationship?
Let's look at how this works out with something like facial attractiveness.