it got worse @andwhatyousaid - Tumblr Blog

Jun 1

Touched for the Very First Time

Lately, I've been having a few conversations about normative assumptions/value frameworks/conventions within fic.

I do want to be as clear as possible, here: I am personally quite hardcore anti-censorship and strongly believe that any attempt to eliminate fics based on ethical evaluation of their narrative content does significantly more harm than good.

With that point established, I also find it absolutely fascinating to examine the ways in which certain patterns emerge! Therefore, when I had the stray thought "I feel like Steddie fandom gets weirder about virginity than other fandoms I've been in," I went—okay, let's test that feeling with data.

It was so, so much fun! I'm having a fantastic time.

Method/Approach

While logged in to my AO3 account (thus with visibility on archive-locked fics), I went through the AO3 ship tags for the most popular ships of 2024, the most recent overview available from @toastystats. Steve Harrington/Eddie Munson didn't make the cut at 23.3k in 2024, but note that that number does not include archive-locked fics.

Within each ship tag, I filtered for (individually):

#Loss of Virginity

#Virginity Kink

#Virgin [character A]

#Virgin [character B]

I did not have a hard cut-off for posting time, but I did grab everything within the span of about an hour.

I decided not to pull rating data (e.g. what percentage of fics in the fandom are labeled Explicit), because I think stories featuring virginity could hypothetically be in just about anything. If I ever do decide to do a follow-up, that's something I'd probably look at, though. Who knows, maybe there'll be some interesting correlations.

I also looked at the stats for "Uchiha Sasuke/Uzumaki Naruto" because Naruto made the biggest fandoms list and I was curious, having sort of nominally been on the fringes of that fandom in the mid-2000s; it's got 21.5k fics so I think it's a pretty reasonably-sized inclusion.

Something I did not expect to discover was the fact that "#Virgin Rey (Star Wars)" kept popping up as one of the top suggested tags whenever I typed in a "#Virgin [character]" tag, so I dipped a cautious toe into the heretofore-unknown-to-me world of Reylo. How could I not, with bait like that?

(Steve Rogers' virginity or lack thereof also seems to be of paramount interest to the good people of AO3...)

Reylo-specific Weirdness

I'm not really sure my approach was sufficient, because I discovered that "Kylo Ren/Rey" (16,748 fics) and "Rey/Ben Solo | Kylo Ren" (31,759 fics) are listed as two separate ships. I did a quick search for any AO3 news items explaining why they're separate and didn't see anything; if anyone knows what's up with that, please let me know! I'm not saying it's wrong, I just have no idea what's going on or if it is indeed appropriate for purposes of this study to treat them as separate.

The numbers I got for those ships are also slightly off, because I decided to sum "#Virgin Kylo Ren" and "#Virgin Ben Solo" to get the "#Virgin [character]" figure, even though I could already see a ton of fics tagged with both on the first page of results. I suspect the raw sums ended up being off by about 100-200, since the lower number in both cases was ~300.

(In both ship tags, "#Virgin Rey" was still the higher of the two sums, so it didn't end up mattering at all in terms of getting a percentage range.)

Key Methodological Weaknesses

⚠️ Obviously, these tags are not comprehensive.

I poked around a bit to see if I could find any other commonly used tags; I considered including "#Sexual Inexperience" and "#First Times" as well, but felt that those were too imprecise and included a lot of e.g. first times between those particular characters. Ultimately, I decided that what I was actually interested in was references to "virginity" as a specific socially constructed concept in and of itself.

⚠️ I did not clean the datasets.

For example, these numbers include fics in which the tags refer to a different ship than the one I filtered by, such as multi-fandom anthologies and works in which the filtered-by ship is a background pairing.

I'm gonna be honest: proper cleaning sounded like a lot of work and I didn't want to do it.

⚠️ I did not account for multi-tagging.

I'm absolutely certain that many if not most of these have more than one tag applied, e.g. "#Loss of Virginity" and "#Virgin [character A]" on the same fic. Therefore, the total number of unique fics is definitely less than the sum of the tags. Which is fine, because I'm interested in looking at tag frequency, but it is a limitation that should be top-of-mind when looking at this data.

Findings

These are raw counts, which means that they're not scaled to the overall fandom sizes as indicated by the grey "Fic total" columns (scaled to right-side Y-axis, not shown).

Given all the listed methodological weaknesses, I want to emphasize yet again that this is an extremely sloppy study with a lot of holes in it, and no firm conclusions should be drawn from it. I know some of the numbers aren't really visible here and honestly I think that's appropriate, because they should be taken as relative indicators/estimates rather than precise measurements. (Actual numbers at end of post.)

Nonetheless, I found these results pretty interesting in terms of overall trends! For my purposes, getting a general range was sufficiently accurate, even if I had a hell of a time figuring out how best to visualize it within Google Sheets.

Here's a chart showing approximately how many times those four tags show up in fics for each ship, described as a minimum percentage (the highest number of any individual tag) and a maximum percentage (the sum of all four numbers). The actual number of unique fics that use at least one of those tags is definitely somewhere between the two, and probably much closer to the minimum than the maximum.

INCREDIBLE. Reylo fics tend to use at least one of those tags at a rate roughly 3-5x any other ship. I have no idea what's going on in there and I'm a little bit afraid to ask. (No, that's a lie, I am so anthropologically interested, please tell me if you know anything!!)

The other major outlier is the Miraculous Ladybug ship, Adrien Agreste/Marinette Dupain-Cheng; I know literally nothing about that fandom/canon and have no idea what's going on there either, but there are only 65 fics tagged "#Loss of Virginity" + one tagged "#Virginity Kink" (which is a multifandom anthology, so I strongly suspect it shouldn't count, but I didn't want to be inconsistent about my hands-off approach to data-cleaning) + one tagged "#virgin marinette."

So with those three (counting the two Reylo as separate) removed to make the chart a bit more legible, I got this:

Second figure included because it's closer to what I actually wanted the data visualization itself to look like but there doesn't seem to be an easy way to add the actual numbers, blech. GIVE ME BOTH.

By this point, you may have noticed that I wrote out some names in full and some in shorthand. This is purely based on my own personal convenience, because it's not like I'm going to forget what LWJ/WWX stands for, but if I shorten the BNHA characters' names I will simply never remember them again. I did, however, write out the character names instead of the ship names e.g. Johnlock because that helps me remember at a glance which one was Character A and which one was Character B.

Main Observations:

Hermione/Draco and Steve/Eddie have a VERY similar spread of around 1.4%–2.4/2.8%.

Jungkook/V (BTS RPF) also overlaps a bit, though with a lower maximum; this may be due to e.g. a fandom tendency not to use multiple of the listed tags in the same fic, though this is just an example hypothesis and further investigation is required to draw any particular conclusions.

Sherlock/John is the only other one with a min% of >1%, though Bakugou Katsuki/Midoriya Izuku (BNHA) and Harry/Louis (1D RPF) come fairly close with a ~1.3% max% for both.

I also noted some interesting patterns around tag frequency/distribution when I was pulling data:

I went with a percentage format for these data markers, but please note yet again that these are not percentages of unique fics! These percentages represent tags used, and the ratio of tag:fic is definitely not 1:1.

Main Observations:

"#Loss of Virginity" tends to be the most or second-most commonly used tag, with "#Virgin [character A/B]" also frequently used.

"#Virginity Kink" is not that commonly used overall, which I find pretty interesting in and of itself! It shows up pretty heavily in LWJ/WWX, which makes sense because they are canonically kinky (I don't remember whether the incense-burners featured virginity kink specifically? but it's definitely not too big a stretch), Dean/Sam (SPN), and of course Steve/Eddie, which has the highest actual number of tag usages (87 compared to the next-highest at 57).

Some fandoms have a pretty even split of which character gets the "#Virgin [character]" tag, and others don't.

On that last point, I wanted to take a closer look at which ships had strong discrepancies and which had a more 50/50 split.

The formula I used was basically in the format of (MAX(A1:B1)-MIN(A1:B1))/SUM(A1:B1)—in other words, the closer the number is to zero, the closer the "#Virgin [character A]" to "#Virgin [character B]" ratio is to 1:1.

I wasn't expecting these results! I'd initially thought that it might make sense that some ships include characters who have unambiguously canonically had sex, so the bulk of canon-verse fics would thus reflect a certain weighting.

However, as the second (sorted) graph shows, below the top two, the curve is pretty smooth! I don't really know what that implies except that my hypothesis was definitely wrong. There does seem to be a more jagged step-down from 0.53 to 0.41, but I'm not convinced the R-squared is out of bounds, especially when n=23.

(I've included the Reylo numbers, but here is a gentle reminder that they should definitely be higher due to fandom-specific tagging practices that I don't know enough about to resolve appropriately. The maximum end of that range would be about 0.3-0.35, for the record.)

(Lack of) Conclusion

⚠️ It would be intellectually irresponsible to come to any overarching conclusions based on this data. (This is mostly a reminder to myself!)

What I can reasonably say is this: the data does in fact seem to support the idea that Steddie fics on AO3 may have more of a tendency to use virginity-related tags than other popular ships do, particularly the "#Virginity Kink" tag.

I am not even going to try to speculate on whatever's going on in the Hermione/Draco situation, but it's utterly fascinating to me that that's the main highwater comp, especially since the overall numbers are relatively similar but the internal tag distribution looks pretty different.

And finally, because I do know it's a pain to try and read the actual numbers on the graphs, here's a cleaner table:

laundrybiscuits

Ok, I went back to edit a few more graphs into this post (I tweaked the transparency on the fic-count columns in the first major graph, so it should be slightly easier to read now) and I realized I was basically starting a whole other very lengthy section on distribution, so. Please consider this a vast expansion on my first set of observations, which in retrospect were quite cursory and inadequate!

Naturally, this is almost as long as the previous post, with even more graphs/charts.

Overall Distributions and Actual Statistics on AO3 Virginity-Tagging Which I Neglected to Include Last Time, i.e. My Formal ADHD Evaluation Came Back Negative Because "Most People With ADHD Find Math Boring"

The tl;dr:

It is unusual for more than ~0.85% of fics to be tagged "#Loss of Virginity" or ~0.05% of fics to be tagged "#Virginity Kink." It is also unusual for more than ~0.45% of fics featuring any particular character to be tagged with "#Virgin [character]," regardless of what ship they're written in.

Here are the initial stat workups:

And here is the general tag distribution:

Note that Reylo and the Miraculous Ladybug ship have already been excluded from these analyses.

What we can see with the consistently negative IQRmin values (Q1 minus 1.5*IQR) (yeah the name's not perfect, it's basically shorthand for the lower bound of a box-and-whisker plot) is that there are no lower-bound outliers in the data.

You will also notice a couple new columns on the initial stats, because I've really been puzzling over whether the character tags should be treated as one continuous dataset. In other words, do I consider "#Virgin Aziraphale" and "#Virgin Crowley" as two points on a single line? Or do I look at the characters within a ship who are more commonly V-tagged as their own set?

Here is a barely-legible distribution map of the more commonly V-tagged (#VMax) and the less commonly V-tagged (#VMin) (yes I'm aware that vmin is a BTS RPF ship name, don't come for me about my naming conventions, it's all just napkin scribbles):

As you can see, there is a fair bit of overlap! At this point, I reeeeeally could not avoid figuring out a proper way to excise outliers, lest my graphs be completely unreadable.

There are a few different ways to calculate outliers, depending on what you're looking to measure, and I decided to test both z-scores and IQRs because that sounded like fun.

Z-scores

Z-scores, which measure against the mean, are typically best used for normal distributions—which these definitely are not. But I just wanted to see what I'd get if I applied it, so here we go.

The usual cutoff is +/-3, which means the only proper outlier on these tags is Steddie on "#Virginity Kink"! 👀

That said, the Hermione/Draco "#Loss of Virginity" z-score is preeeetty close to 3, so I'd be comfortable considering it an outlier in this context. (A former stats prof once described the process of finding outliers as "vibes-based" which probably explains a lot about what I'm doing here)

Now for the character tags:

The results are in fact different depending on whether I use one dataset or split it into two, which is SO COOL!

What the above chart means, basically, is that when each individual character is considered equally hypothetically likely to be V-tagged independent of who they're shipped with, Eddie Munson and Sherlock Holmes are VERY disproportionately V-tagged. The rest of the characters are pretty tightly clustered within about +/-1, except Hermione Granger at 1.80.

However, if we make the assumption that any given ship is likely to have one character V-tagged more than the other*, Eddie is flirting with that upper bound there but only Draco Malfoy really breaks it—as the less-tagged partner.

*Should probably have mentioned this earlier but all of the top listed ships have 2 characters, so I'm not considering polyships as part of this analysis.

Also, while Steve Harrington is slightly more likely to be V-tagged than most characters on the VMin side, John Watson is slightly less. Neither breaks +/-1, which I hadn't expected!

Interquartile Ranges (IQRs)

As mentioned earlier, the lower-bound IQR can be ignored as we don't really have outliers on that end. Here are the upper-bound outliers, sorted by min%:

Again, we do get slightly different results depending on whether the characters are treated as individual or as participating in an oppositional binary, but this time the difference is pretty trivial.

To summarize the non-character tags first: Hermione/Draco has a slightly higher percentage of "#Loss of Virginity" tags than most (Jungkook/V is also creeping up there); LWJ/WWX has a slightly higher percentages of "#Virginity Kink" than most (with Dean/Sam literally right on that line); and Steve/Eddie has about 2.5x the "#Virginity Kink" IQRmax, which I would classify as EXTREMELY HIGH, HOLY SHIT.

As for the character tags: unsurprisingly, the top three VMax entries on this list all clock in as outliers. However, the only real difference is that Draco is marked as an outlier for a VMin.

Conclusion Regarding Outliers

Given all of the above results, I think it's pretty fair to say that Hermione/Draco, Steve/Eddie, and Sherlock/John are outliers adn should not have been counted.

So what happens if we remove them from the distribution? (As well as the two other "#Virginity Kink" outliers)

Now that looks like a pretty reasonable distribution all around. The n is rather too low to expect normality, but it does seem like these could be representative sample populations. (I'm not even including the exact data labels because these specific charts were created to check for clustering or other weirdness, and thus the exact data labels don't matter.)

In the end, I did write out some of the characters' full names and marked which ships they belong to so I could make more detailed comparative charts. As you can see in the above charts, there is a TON of overlap.

(Note that Steve Rogers is V-tagged at approximately the same rate whether he's paired with Bucky or Tony, which I personally think is hilarious! Fandom consensus, truly. I'm slightly curious whether comics!Cap is treated differently from MCU!Cap, but that is a project for another day.)

Statistical Summary Revisited

With the outliers removed, this is the statistical summary I came up with. I'd like to test it against bigger datasets, but it feels about right to me.

I feel fairly confident in saying that based on this data/analysis, most popular ships feature 0.2-0.6% fics tagged "#Loss of Virginity" and 0.01-0.03% fics tagged "#Virginity Kink."

It also appears to be completely within the bounds of a normal distribution to have zero fics with virginity-related tags, which means I should maybe have left the Miraculous Ladybug ship in, whoops.

And here we get to the actual, real conclusion of this extremely long journey! If you forgot the tl;dr before the cut, here it is again:

It is unusual for more than ~0.85% of fics to be tagged "#Loss of Virginity" or ~0.05% of fics to be tagged "#Virginity Kink."

It is also unusual for more than ~0.45% of fics featuring any particular character to be tagged with "#Virgin [character]," regardless of what ship they're written in.

So back to my animating research question...

Is Steddie fandom relatively weird about virginity?

As I've said many times throughout these posts, I can't and shouldn't make any conclusive statements based on this data.

Nonetheless, the data I've examined has not disproven my hypothesis. And that's not nothing! It's a bunch of double negatives, sure, but sometimes that's what you get.

Steddie joins Johnlock and Dramione in this presumptive preoccupation with virginity, although it's much more similar to Johnlock in that it is primarily concerned with the virginity of one character (a twenty-year-old musician and drug dealer) in particular. Steddie fandom is, however, MUCH more likely to tag it as kink, and is actually pretty normal about tagging "#Loss of Virginity." "#Virgin Steve Harrington" also shows up on the sliiightly high side of average, but is actually also very normal.

Potential Future Research Directions

First and foremost, it would be cool to check these stats against a bigger dataset; maybe the top 100 ships or something like that.

I'd like to check the co-tagging/multi-tagging: what's the prevalence of fics with both "#Virgin [character A]" and "#Virgin [character B]" tagged? What are the most common tag combos?

It would also be interesting to do a comparison of M/M and M/F ships (and F/F, I suppose, but I strongly suspect F/F ships tend to be less attached to the concept of virginity—though maybe that's another assumption worth testing). I was actually quite surprised to see that both Reylo and Dramione have relatively low discrepancies between "Virgin [character]" tags, and I'm curious whether that's characteristic of M/F ships in general.

I'd like to see how "#Anal Virgin [character]" tags play out relative to the ones I checked. They did pop up quite a bit in my data-gathering, but I decided they were out of bounds for this particular study.

I'm curious about two other tags I'd decided were out of bounds, too:"#Sexual Inexperience" and "#First Times." I do wonder if they change the numbers at all!

I have no idea whether this would be fruitful or not, but it'd be neat to check for any potential correlations (or inverse correlations) with omegaverse tags.

This is a bit tougher to study, but I'd like to investigate cultural differences around fics tagged for virginity. One imperfect proxy would be using the language the fic was published in, but I feel like I'd need a couple other variables to really start getting meaningful data. Maybe looking at e.g. C-dramas in their own category?

WHAT IS GOING ON WITH REYLO

#obsessed #thank you truly for your fandom contributions laundrybiscuits; there is only one of you #and i wish we could pay you full-time