How Relationships Change Over Time

Survey results

I asked ~14,000 people, 8739 male and 5230 female (though this sample might increase as I write this survey/update charts), about their romantic relationships.

If you’re in a relationship and haven’t taken the survey yet, please do it now before you read this article and get primed!

Survey sources were from my social media, and people they decided to share the survey with (I designed the survey to give results, which increased virality a bit). My friend, who runs a large website that does personality testing, posted a link to my survey there as well (and can compare difference in responses from my source vs. the testing source).

I’ll talk more about the kind of populations I was sampling at the end of this article, but before you see the results, I’d like you to preregister your guesses - given what you know about the nature of the survey and who took it, what would you predict the results will be?

I’m presenting the data without adjusting for anything; you can doublecheck my work if you’d like, I’m still learning as I go, and your feedback and error-catching is very helpful to me! Data is here, but right now I’ve deleted a lot of columns that are potentially identifying; once I look more deeply into ensuring fully anonymous data then I’ll share more. My python code is here; I’m a very noob programmer and I apologize in advance)

A few points - I’m not separating out trans and cis or gay and straight - this is just a straight split by biological sex, and it’s possible you might see very different results if you split by identity or orientation. If my sample is big enough I might write more on that later!

Also, some questions were asked as negations, and I put those in parenthesis - for example if you see “I (do not) have good sex”, the question was asked to test takers as “I do have good sex”, and I just flipped the sign and the question for the purpose of clarity in the graphs.

And as a last reminder - this is asking people in currently existing relationships. We don’t know if the graphs change because people are dropping out of relationships and thus the sample, or if relationships are actually changing over time - or both!

Frequency of sex over relationship length:

Average size of each bin of length amount was n=939.

I am pretty impressed that male and female ratings seem to be nearly identical here.

Frequency of fighting over relationship length

Scores

Before I ran this survey, I had around 600 people answer around 80 questions about their relationship, and I did a factor analysis on their answers with oblimin rotation. I don’t really understand factor analysis deeply, and I’m not sure if this was the right thing to do, but this is what some data people recommended to me. I wasn’t trying to do an extremely rigorous thing, mostly just wanted to get a few measures on which to glimpse how a relationship is doing, so I could ask people about it on my follow-up survey.

I took the top 7 factors, which I’m naming health, toxicity/fighting, insecurity/jealousy, codependence/interdependence, friendship, openness, and sex. I’ll give you the exact questions as I go through.

Health

The health score is something I interpret as non-codependent flourishing; a solid, beneficial base that allows you to be a better person. The score is the average of answers to the following questions:

"I am careful to maintain a personal identity separate from my partner"

"I think this relationship will last for a very long time"

"This relationship is good for me"

"My partner enables me to learn and grow"

“I feel confident in my relationship's ability to withstand everything life has to throw at us"

(negative) "In hindsight, getting into this relationship was a bad idea"

People rated the questions (and all the following) on a -3 to 3 scale, from “strongly disagree” to “strongly agree”, with 0 as neutral.

It looks like women, on average, rate their relationships slightly better than men no matter how long the relationship. I’m also surprised to find that the ratings are mostly stable over time!

I wonder if this is some artifact of smoothing of the individual questions though? Let’s see:

Kinda unclear! But after looking at this, I think I’m just going to present the rest of the data as granular questions, though I’ll still group them by the factors. I’ll split it out by male/female if there seem to be any noticeable gaps; otherwise they’re probably pretty similar.

Squinting, it looks like the quality of relationships seem to take a dip around 7-8 years. I’m a little surprised to see the ratings so stable in the very early stages of a relationship!

Toxicity

I asked 6 questions about toxicity. It gets a bit complicated when I try to fit all the info into one graph, so I’m splitting them into 3 each.

It seems men I surveyed find their relationships more toxic than the women I surveyed; women’s average ‘toxicity’ score was -1.6, while men’s was =1.35.

The little dip, mostly for men, at 6-9 months is interesting to me. Is it just noisy bins from low sample size? The count in each bin (male and female combined) was 677, 795, 545, 913, 959, 792, 1353, 1107, 1258, 683, 670, 878, 1187, 954, 1198 - all pretty big bins, so I don’t think so?

Maybe 6-9 months is a pivotal decision time for relationships; people will generally tolerate a bad relationship, but at that point if the relationship is bad, they drop it - thus we see them drop out of the sample and the remaining 9-12 month relationships report having a better time?

Not sure; it’s a pretty small dip anyway.

I am a little surprised to see relationship toxicity remaining mostly stable, then rising faster at around 7 years into the relationship. I would have guessed that if a relationship was gonna get annoying, it would hit much sooner. Isn’t 4 years plenty of time to get sick of your partner?

Is this kid related? Is there a certain age of children that makes things terrible?

Males with children reported having their first child on average 6 years into their relationship, and females reported 5.3 years. This might line up, assuming kids start to get really annoying once they’re 2 years old or so? Or maybe this is around the time the second child gets born, and two kids are much more difficult than one?

Let me spot check a few questions here:

Lighter lines have kids, darker lines don’t have kids, and I didn’t have much of a sample for very-early-in-relationship kids, so I averaged a few of the lower bins and started late.

Looks like people with kids rate the judgment and relationship fights as worse - a whopping 0.5 points on average worse (reminder, this is a -3 to 3 spectrum). It’s unclear if it makes relationships even worse over time - I think I need a higher sample size of more people in longer relationships for that.

But compare the line indicating “We handle fights well” a few graphs above, to the one of people without kids. When I check the data, people who have kids rank their ability to fight well at 0.95; people without rank their fight ability at 1.38 (tho reminder I’m still not controlling for anything with these numbers).

Sex

I asked an absolute ‘how often do you have sex’ question, but I also asked questions about sex-related satisfaction; passion, excitement, desire, etc.

Since sex is more gendered, I’m gonna give you a bunch of graphs broken down by gender, too.

(spoiler: they’re all saying almost exactly the same thing)

It looks like women in general are a bit more satisfied with their sexual relationship than men, but that the gap grows over length of relationship.

I’m again curious about kids - let’s test one of the questions.

From glancing at this graph, it doesn’t look like there’s an obviously big impact of kids on sex life. Still though, bins were pretty low - smallest bin was females with kids in their relationship 3-4 years, with n=21.

Interdependence

Or maybe codependence? Depends on how you mean it; these questions were about intertwinedness of life, dependence on each other, identity wrapped up in the other, etc.

Relationship interdependency rises pretty solidly over time, with the “self worth” question remaining stable and the “if my relationship ended I’d be devastated” question taking a dip around that ominous 7 year mark.

It also seems like the most rapid climb happens in the first 9 months; after that the increase interdependency still climbs, but more slowly.

Insecurity/jealousy

A nearly universal decrease in the jealousy questions over time! Pretty cool. I also like how it seems like the decrease really starts dropping after around 1 year; maybe very jealous relationships tend to explode around this point, or maybe it takes around a year for people to feel reassured, on a primal level, that their partner won’t stray.

Let’s spotcheck this with gender:

Looks like a pretty similar change over time for both males and females, but females overall report higher wariness about other hoes stealing their man.

Friendship

Friendship-related questions seem remarkably stable, with a hint of that slight ominous dip at around the 7 year mark again.

Out of morbid curiosity, I checked the average ratings for male and female on the ‘intellectually matched’ question; females agreed with the question “we are intellectually matched” at 1.73, and males at 1.22.

Openness

I just realized I could do dashed lines in google sheets graphs!

Females seem to have fewer secrets, share more, and explicitly discuss the relationship more than males, and it seems like this gap grows ever so slightly over time.

And again, things start to start sliding off around the 7 year mark.

---

For this post, I’m trying to focus this post on the type of questions that put length in the x axis of graphs - there’s more questions I’ll do more writeups in the future! But for a few closing things to check:

This chart is so clean I want to cry. Looks like marriages start around 1.5-2 years into a relationship and then very steadily increase.

I also asked how much time were you in a relationship before you got married? This graph is married people only:

To clarify this graph, example: of married people who are 1-1.5 years into a relationship, they spent 10 months in that relationship before getting married.

I’m not sure how to interpret this, but it might be some evidence that roughly 4.75 years of dating (but not more) before marriage corresponds with longer marriages. To walk this through slowly in a very dumb and probably obvious way:

if you have a bunch of married people who’ve been in a relationship for 8 years, and a bunch of people in a relationship for 11 years, and you ask them all how long they dated before getting married, the 8-year group will report ~3.8 years, and the 11 year group will report ~4.75 years. My guess here is that we’re actually seeing some effect where as the 8-year group ages, the shorter-daters disproportionately have relationships fail, and by the time they hit 11 years of relationship the remaining ones will be the ones who dated longer. This doesn’t seem to be true beyond this - 11 year relationships and 16 year relationships all have roughly the same amount of pre-marriage dating time; as in, whatever makes relationships fall off after this doesn’t seem to correlate with how long you spent dating.

This specific type of question feels like unusually hard for me to think about - am I making sense? I’m trying to disentangle “of course you’re not going to have any 4-year pre-marriage dating periods in a 3-year relationship” from seeing relationships actually falling off the chart and only the successful ones left over.

Anyway, moving on:

Well would you look at how the corner of that curve seems to be about 7 years.

On a “how religious are you” scale from 0 (not religious) to 4 (devoutly), it seems mostly stable across time. This is kinda good for me - it means I’m getting a mostly nonreligious sample across the board (as opposed to spots of religiousness that might end up being a confounder).

External intimacy interest

I finally decided to limit this to cis, hetero relationships, since the sex of the person’s partner probably matters a lot more for this question.

The verbatim questions for this were

“In a world where your partner was fully aware and deeply okay with it, how much would you be interested in having sexual/romantic experiences with people besides your partner?”

“In a world where you were fully aware and deeply okay with it, how much would your partner be interested in having sexual/romantic experiences with people besides you?”

And the response options were “not at all” (0), “a little bit” (1), “somewhat” (2), “a lot” (3), and “extremely, this is essential” (4).

Both males and females predict their partners would be interested in extra-relationship intimacy with a score at around 1 (which corresponded to “a little bit” in the survey), with this number increasing very slightly over time (total female respondents in my survey, not controlling for anything, averaged 0.98; males averaged 1.02).

Females do in fact report slightly higher interest in extra-relationship intimacy themselves than what males predict, especially around the 3-4 year mark - but the discrepancy in how females predict their male partners (avg of 0.9) feel vs how males answered this survey (1.8) was much bigger.

Cheating

I asked people “Have you or your partner ever cheated on each other?” with the answer options “No”, “Yes, I have”, “Yes, they have”, “Yes, we both have”.

Or for a more fun little comparison:

Or, even more fun, let’s limit this to cis, hetero relationships again, and see the cheating people know about vs. the cheating they admit to:

Kinda brutal, no? Females seem to have a much smaller gap between actual cheating and perceived cheating than males do; over 1/3rd the males who said they’d been in a relationship longer than 16 years, reported cheating on their partners!

Porn

There was an option for “more than once a day”, which was weighted at below 1, so this graph isn’t strictly a representation of “days between porn watching", which might imply multiple porn sessions in one day count as one.

Men are incredibly stable, and I wouldn’t be surprised if at least some of their slight decline was due to age instead of the relationship. Women drop their porn use a lot over time, their averages nearly halving at the extremes.

Demographics

What kind of population was I sampling? Pretty white, liberal, middle class.

Some basic info on the survey respondents:

Average female age was 29.2, male was 33.2.

Average female relationship length was 5.1, males was 7.1.

93% of people identified as cisgender.

Average female reported income was 64k, average male was 108k.

75.9% were white, 8.5% were hispanic/latino, 7.2% were asian, 4% were ‘other’, and 1.6% were black.

On economic and social politics questions, on a scale from very conservative (-3) to very liberal (3), females scored 1.75 on economic liberalism and 2.2 on social liberalism; males scored 0.83 on economic liberalism and 1.7 on social liberalism.

54% came from twitter, 22% came from ‘other’ (which is probably mostly my friend sharing the survey on his personality testing site), 13% came from reddit, 5% from facebook, 3% from fetlife, 1% from telegram, and the rest of the sources were negligible.

People who came from twitter were older (32), more male (68%); people who came from ‘other’ source were younger (28) and less male (40%). This might end up leading to some skew in the data, as it’s possible that whatever process is selecting for age and gender in the sources is also selecting for other emotional traits that might impact things.

I gave people a reward for finishing the survey - information about how they rank - so this hopefully contributed a little bit to it being shared throughout networks.

---

Summary: If I had to squint and do some narrativizing, It seems there’s two mild points where relationships seem to shift - around 1 year in, and 7 years in. It takes about one year to ‘calm down and get secure’ - for the sex magic to finally begin to start wearing off, for orienting around jealousy to get sorted out, for you to know what your relationship is about. Relationships seem to be mostly stable after this, until 7 years hits and the doom begins to set in. People conceal more, cheat more, fight worse. It seems pretty plausible this is kid-related; people have kids, and then their relationship begins to suffer.

(Of course, remember we’re looking at surviving relationships only here. The shorter the relationship in the sample, the more that sample includes relationships that won’t survive in the future.)

I’m hoping in the future I’ll be able to get an increased sample size for people in much longer relationships, so we can track stuff like “do relationships get better after kids move out?”