Who tops more, Draco or Harry?

[snowgall]

Yesterday I was having a discussion with vaysh about who tops more in fanfic, Draco or Harry (link here). I was of the opinion that it seemed fairly balanced, but that was from my perspective of mostly reading recent fic. Vaysh has been reading Drarry fanfiction for much longer than me, however, and she disagreed:

"statistical evidence is very clear: There is in no way a 50:50 balance between top and bottom!Draco fanworks. For many people, bottom!Draco remains the default."

Aha, I thought, a statistical challenge! As it turns out, I have a database of Drarry fics which I can consult, so I am in position to offer an answer to this pressing question. There are nearly 1500 Drarry fics written between 2002 and 2015 for which I have information about who tops or bottoms, and according to this data, we're both right :)

1. Who bottoms more across all fics in all years?
   Answer: Draco, clearly (point vaysh)


2. Has the proportion of bottom!Draco to bottom!Harry begun to balance out more in recent years?
   Answer: Yes, clearly (point snowgall)

Before I get to the fancy chart and graphs, let me explain a bit where my data comes from and how reliable and/or biased it could be.

My list of Drarry fics includes not just fics that I have personally read, but also every fic that capitu has ever recced, as well as every fic ever recced at hd_storyroom, whether I've read them or not. Capitu (at my_drarry_recs) helpfully includes tags indicating who bottoms in a fic (if applicable) so I have that data even for fics I haven't read. I have also included many (but not all) of the fics recced by gracerene, who also includes top/bottom info. The chart below only contains data from fics where I could determine something about who tops or bottoms.

Of course, for fics not recced by capitu or hd-storyroom, which fics get listed in my database will be skewed by my own reading preferences. However, it's important to note that I personally don't care who tops or bottoms, so I don't think that I will have unintentionally skewed the data one way or the other.

The other big issue concerns the date a fic was published. The vast majority of the fics in my database are fics written since 2007, and precious few written before 2005. But I also should point out that if I can't accurately date a fic, then I'm not including it in this dataset. The earlier a fic was written, the harder it can be to find its pub date, so part of the reason there are so few pre-2005 fics is because I can't be sure when they were actually written.

Ok, enough caveats about the data, on to the chart! I hope it is self-explanatory, because I'm not going to waste more words explaining something you can see for yourself :) Just note that I'm including fics from 2015 too, even though we're only 3 months in.

Year	bottom!Draco	bottom!Harry	switching	no penetration	Total
2015	9	8	3	27	47
2014	70	70	26	85	251
2013	97	55	16	64	232
2012	67	57	8	29	161
2011	49	33	15	42	139
2010	54	36	12	32	134
2009	58	26	11	31	126
2008	75	32	7	20	134
2007	54	22	14	16	106
2006	27	11	10	9	57
2005	26	10	14	15	65
2004	11	4	2	4	21
2003	7	3	2	2	14
2002	0	1	0	2	3
Overall	606	368	140	376	1490

 

And now for some pretty pie charts:

 

 

 

 

 

As you can see, before 2010 bottom!Draco was much more common than bottom!Harry, but since then the proportion of bottom!Draco fics has gone down, to the point that of the works I've analyzed for 2014, there's a perfect balance between the two!

Is the influence of the dracotops_harry fest part of the reason for this change? It's possible. The DTH fest began in 2011, and my data seems to date the gradual shift to 2010.

It's also true that some of the disparity could be traced to a few big-name authors who tend(ed) to prefer one dynamic over another. Here's just a few examples I know of authors who seemed to prefer bottom!Draco but who haven't written as much in the last few years:

• ladyvader, active approx 2003 - 2010
• jennavere, active approx 2004 - 2006
• pir8fancier, active approx 2004 - 2013
• silentauror, active approx 2005 - 2009

There are many more bottom!Draco authors who are still actively writing HD, among them: enchanted_jae, oldenuf2nb, and samaelthekind.

But interestingly, I can only think of one long-time author who seems to prefer bottom!Harry: megyal.
(Edit: vaysh pointed out that lomonaaeren is also a bottom!Harry author. My bad!)

If you have any insights into this data, disagree about my analysis, or can think of something I missed, please leave a comment!

 

---

 

Edit: Inspired by suggestions in the comments section, I have re-run the numbers using just data from the hd_fan_fair fests. See my write-up here. The data was collected through crowd-sourcing, and we are currently working on building a second crowd-sourced spreadsheet of data from hd_smoochfest too!

Comments

[torino10154]

I saw this linked two different places and was intrigued. I'm only peripherally in the H/D corner of fandom, in that I write them sometimes but don't read a whole lot. As a writer, I think I default to top!Harry but I just did an mpreg fic where it was Harry pregnant. *shrugs* I'm one of those that lets the story lead me and don't have a strong preference going in. Oh, and I do like to imply switching even if in the fic itself only includes one act of penetration.

From this position, I'd say that I would have assumed there would be more bottom!Draco over all but that yes, things are evening out. Not equal, no, but not as lopsided. I think the fact that there *is* a Draco Tops Harry fest should tell you, though, that there was a market for something that wasn't the norm. Meaning if there was a Harry Tops Draco fest, wouldn't people find that unnecessary? Maybe there will come a day when the pendulum swings far enough that's the case, but as the stats seem to indicate, top!Harry is still the majority.

Bringing a bit of insight from Severus/Harry, where I do spend my time, there is a bottom!Snape community precisely because he usually tops. People interested in bottom!Harry don't have to do anything to find it. And, unlike H/D, I think that's still the norm. Perhaps because of the age difference (and common assumption that older=top).

tl;dr ramble LOL Thanks for the interesting stats!

[vaysh]

Storm Constantine fan here, indeed. :) Vaysh is one of my favourite characters.

[snowgall]

You're welcome, and thanks for the information about the Harry/Snape dynamic, which I knew nothing about!
Assumptions about who should top or bottom based on canon characteristics always make me laugh, though, because canonically they're both straight! I mean, I like my fanfic characters to be depicted as close to canon as possible, but I don't pretend that this says anything definitive about who would top :) And I'd imagine that if someone liked to bottom, that wouldn't necessarily change as they got older!
(This is not a complaint against you at all; I know you were just speculating on what other people may be thinking. I'm just musing aloud and rambling too :)
Oh, and there is a bottom-draco fest, although it's not HD specific.

[vaysh]

Do you know gossymer's list of fandom classics (up to 2007)? I am sure some of the links are dead by now, but this is a pretty comprehensive list of well-loved H/D fics from earlier fandom. It would take a bit of reading and research to determine which of the fics are top!Harry, which top!Draco and which are switching fics or don't feature penetration at all. But I think, using this list, you could get at meaningful data from early fandom days. :)

[vaysh]

As for the two authors: Maya's fic very rarely ventured into R-rated territory, and I don't think she ever described graphic penetration in those few sex scenes. I remember trying to figure out whether Harry tops Draco (Harry tops with another man in the story) in Drop Dead Gorgeous, and the sex scene is done in such a skillful way that it's impossible to say. :)

Calanthe, on the other hand, is a rather graphic writer. I think the majority of her fics are bottom!Draco, with notable exceptions (Bad Blood, e.g.). I have them saved, and can check, if you're interested.

[eidheann_writes]

I don't know much about data sets, but would it be more telling to look at maybe a specific fest over the years (hd_fan_fair for example, which is a prompt (not a gift) fest which is long running but less likely to include people feeling obligated to write topping preferences due to prompts rather than gifts) and then use those stats over the years to draw conclusions?

I mean, smaller sample size but less chance of unintended bias?

[snowgall]

Oh, very interesting to know about Maya. And I've read a few of Calanthe's fics but not all, so any info you have would be appreciated. Especially if you know anything about dates! Because if I have the top/bottom info but no date, I can't as easily average it with this data (other than for the overall picture, not for the year-by-year). I'm still interested in whatever you can tell me though.

[snowgall]

I think I went through that list at one point and pulled out all the fics that looked interesting to me. I may go back and grab a few more, but if I'm not interested in reading them for pleasure then I don't think I'm going to read them for work ;) If anyone has top/bottom info on early fics (especially ones not in capitu's or gracerene's lists, since they give top/bottom info) I'm very happy to add it to my database. I just already have a to-read list that has hundreds of fics on it, so even if I want to read more it won't happen very quickly!

[snowgall]

That's definitely another way to go about it. I have a number of fics from fan-fair in my database but not all. I happen to know that between 2010 and 2014 there were 157 fics posted to the fan-fair fests, but I only have top/bottom info for 78 of them (the ones capitu recced and/or I read). So that still leaves 79 fics that somebody will have to skim/read to find out the info for, unless somebody somewhere already has it? And my to-read list even without those fics already has humdreds of fics on it, so I don't have enough time to read 79 more :(

So that's what it all boils down to in the end. It could be that one set of data would be more representative or fairly balanced than another, but unless someone takes the time to extract the data (by reading each fic) then there's no way to do an analysis.

There are too many freaking fics! But really, this is a good problem to have in a fandom :)

[snowgall]

OK, let's try this again for #4 :)

So as for percentages vs. ratios, I'm not sure I see what would be gained. In particular, the ratio of top!Harry to bottom!Harry fics *should* just be the reciprocal of the ratio of top!Draco to bottom!Draco fics. This is because top!Harry = bottom!Draco and top!Draco = bottom!Harry. If you're getting different ratio reciprocals, it'll be because of wonky tagging, and I don't think that tells you anything special.

Or am I missing something? I admit that while I'm a math person (I have an undergrad degree in math), I'm not really formally trained in statistics, so if there's something I don't understand please tell me. I don't mind being schooled :)

[snowgall]

#5: how to get a true random sampling.

Christ, I don't know :) Eidheann had an idea below about just looking at the stats for one fest over several years, and that could be interesting, but it still mostly limits you to people familiar with LJ. And also it would mean me having to go a read a bunch more fics that I hadn't planned to :)

Ultimately, unless there is a magical way to create a database of every HD fic ever written by anyone on any platform, there's always going to be some inherent bias in any given subset. We just have to do the best we can with what we have.

I readily admit that the data I presented is not the definitive word on the subject, and in fact there could well be a few errors in my database. Indeed I already suspect that some of my judgment calls as to which fics are one-on-bottom and which are switching could be debatable, but I think that would only be for a small handful of fics and not really affect the big picture in the end.

And it's ALSO possible that I have some of the dates wrong. I spent A LOT of time going through and trying to find the first instance of a fic, but I might not have always gotten it right. For example, if a fic was posted to AO3 in 2013, say, but was really posted originally to some other blog/archive in like 2010, it's possible I didn't find that and so still have it (incorrectly) listed as 2013. I did my very best, but I'm only human. Whenever I find an error I go back and fix it, so my database is always a work in progress.

This makes me think it will be fun to revisit this question again in like 6 months or a year, when I will have read even more fics, and probably fixed more errors, and we can see if anything has changed :)

[snowgall]

And now the last question, #6: is there more non-penetrative activity these days than there used to be?

Well, to answer this would be a whole 'nother blog post in the making :) It is fortunate that as I read, I do make additional notes about what kind of sex goes on if it's not penetrative, but I only have that data for fics that I personally have read. Capitu and gracerene don't tell us that (nor should they!) This means that this dataset is necessarily much smaller than the info I have on top/bottom. PLUS, I wasn't so good at taking notes in the beginning, and only got more detailed the more I read and the bigger my database got. So for many of the fics I read back when I was a newbie, I don't have that data :(

And does it count if the sex is "offscreen" or only implied? Is it still top/bottom then? Vaysh asked that somewhere above, and really, no one is going to agree about this either!

So, I hope that answers most of your questions :) Please let me know what you think - your questions were really good and made me do a lot of critical thinking ♥

[dicta-contrion.livejournal.com]

YAY these answers are amazing! I, too, could get essay-esque on this but will try to keep it brief because, long day, lots to address.

On popularity - agree that kudos are an imperfect measure of popularity, but would be surprised if what happened with "Bond" is a common story, vs. new works being uploaded. And other measures of popularity - how many times a story is recced, for instance - are hard to keep track of, especially once we get into what does or doesn't count as a rec (reblogs? retweets? it would be nigh-impossible to settle it, let along count). But if a majority or strong plurality of fics have less than 5 kudos and your sample has 3, that doesn't sound representative to me.

What makes me curious about popularity, though, isn't representatitiveness qua represntativeness, but whether the stories that are unpopular have anything in common. Are they unpopular because they have loads of SPaG issues or poor characterization or are unusual in some ways? Possible. But I wonder - don't hypothesize, but wonder - whether there's something content-related at work. So, is it possible that some stories are unpopular because of their content? Such that excluding unpopular stories would also exclude content that would be represented by a random sample?

(more subsequently!)

[dicta-contrion.livejournal.com]

Right! Your numbers make sense. I was drawing from your list of capitu's fics and cutting duplicates, but the publicly available author-alpha list doesn't have ratings, so that 444 figure includes authors who wrote smut-free works. So I've no doubt that your 208 or 191 figures are more accurate. We could split the difference and say 200 unique authors.

I don't think there's a magic number of authors that would automatically be enough to prevent weighting towards one person's preferences; it's more to do with how much of the sample they represent, so a percentage issue rather than a numeric one. So perhaps it's not a question of adding more authors, but of whether the results change if you limit each author to one fic (selecting at random which fic will be representative of their work)?

[dicta-contrion.livejournal.com]

Maya and Calanthe are interesting authors to talk about, I think, because while their work is "no longer available on the internet" it's among the most sought-after in my experience. I get requests on tumblr to share works by both authors with some regularity, and while I don't share it, (a) I do have it and went out of my way to find/save it; (b) I think it's telling that there's such tenacious demand for it. And telling, even, that they came up so readily as for-instances.

Does it matter and if so, why? I think yes, but not because they would have made much of a difference in the results. Rather, it's because they probably make a difference in readers' perceptions of fandom - because they're classics, because they're in demand, because they're held up as some not-quite-platonic-ideal of what H/D fic was before canon was closed. And, when those readers become writers, it may also have an effect on how they write the dynamic or how they think the dynamic *should* be written.

So perhaps another metric is relevant: not just the number of authors or number of fics, but the number of views. In terms of our perception of what is/was going on in fandom, and what is/was in demand in fandom, how much of it really is influenced by a few classics? How significant is it to talk about how many top!Harry fics there are, vs how many times top!Harry fics were read?

(In the broader category of "thoughts on this," older fic reads differently to me in so many ways that topping/bottoming is kind of the least of it. But that's a whole other story.)

[snowgall]

If we hypothetically limited each author to one fic in the dataset, then our dataset will no longer be representative either.

It's like the difference between the US Senate and the House of Representatives. (I'm guessing you're an American?) The Senate has two members for each of the 50 states, which might seem fair (like one fic for each author) but then the people in the sparsely-populated Wyoming have much more representation than people in highly-populated California. So not really fair. How fair you think it is probably depends on where you live!

If an author is prolific and writes a lot of bottom!Draco, it will indeed make it seem like there's a lot of bottom!Draco fics out there, *because there are*. So unequal representation is not a bug but a feature.

This might also be a (partial) answer to your question of unpopular fics. If the fics are *truly* unpopular, that is, not many people have read them or ever will, then those fics will not be influencing what people are getting in their reading experience, which is what we're trying to determine, right? That way of looking at it probably seems a bit harsh. I'm not trying to say those fics are less worthy. Just that if people aren't reading them, then they haven't affected reader experience at all.

Of course, I'm a reader so I tend to look at this from the reader's perspective. You are both a reader *and* a writer, so your perspective is probably different. I have a few more thoughts on popularity I'll add in a bit.

[snowgall]

When you talk about number of views, you're hitting on the point I was just making (in response to another of your responses, oy!) that popularity does in fact matter, at least in terms of reader experience :)

[dicta-contrion.livejournal.com]

Oooh, fascinating!! This sounds smart to me on several dimensions. Hadn't thought of doing it like that, but it makes some sense.

(I do wonder if there's a difference between authors who participate in fests and those who don't, and can speculate about what it might be, but I don't think there's anything clear enough there to be convincing. So this sounds to me like by far the best mix of representative and doable.)

[dicta-contrion.livejournal.com]

Now debating whether I'm actually that curious. Especially since one would just have to read the smutty bits.

ETA: Didn't actually mean that in a salacious way! But in a "hey, that's less reading to do" kind of way.

[dicta-contrion.livejournal.com]

Nope, this is completely fair. But it does make me curious about why - when you're right that the ratios should be the same - the tagging doesn't reflect that. Which, I don't think wonky tagging will reflect anything that's quantitatively significant, but might tell an interesting qualitative story.

[snowgall]

So the popularity thing is going to be something that's very hard to settle, but here's a few more of my thoughts on the matter:

a) if a fic has loads of SPaG issues and problems of that sort, then I really can't find it in myself to care very much if it gets counted in these stats. Basically, if an author can't be bothered to run a basic spell check, then I'm not going to make the effort of entering its data into my database. Entering data takes time. I'm willing to do it even for fics I haven't read - and even for fics I've read but not liked - because I like charts and tables and stats, but there has to be some baseline of *objective* quality, even if it's low. Not all data is created equal, in my opinion :)

b) the content issue is more interesting, but also extremely difficult to assess, if it's even possible at all. I just know that from my viewpoint, content varies widely among even popular fics. There are people who like mpreg, and people who don't, people who love poetic, dream-like writing styles and people who don't, people who like experimental styles like non-linear storytelling and those that don't, etc. etc. If there's actually something that somehow bonds all HD readers together in a collective mutual opinion of "we don't like that!" then maybe those fics don't need to be counted, you know?

[snowgall]

If you truly want to do this, then I will happily set up a google spreadsheet for us to share and input all the data I currently have from fan-fair and let you input the rest. Then we can do a joint report on it. What do you think?

[dicta-contrion.livejournal.com]

So, I happen to be sitting here with an HP fan who is formally trained in stats and related methodology. Her supposition is that constructing a good sample would have to involve a lot more randomness. So, for instance, going on all the major archives and taking 5% of all NC-17/M H/D fics and analyzing content.

(Statistician friend also adds that she would be interested to see if there's a difference by venue. She reads primarily on ff.net and in her experience they either switch or Harry tops. I read on AO3 and would say they mostly switch or do other stuff (but I also dgaf and can't be arsed to remember most of the time))
(Also re: the time involved in actual reading, I wonder if one could use python or other content analysis software, though that's sadly just above my methodological paygrade. Who knew this is why I'd want to have made the investment? Hmph.)

But - doing that would take a tremendous amount of time, and it would be incredibly difficult to parse by year and establish chronological trends, so. That.

Also, so important to say, and I realize I haven't said it - this data is tremendously impressive!! Especially where dating is concerned. Avoid errors would be impossible given the information you have to work with, and the time you've put into the dating is a gift. Seeing the chronological shift is among the most fascinating pieces of this to me. I could spend aaaaaagggggeeeees speculating about why that is. Among other things. Kind of like I'm doing. So, this data is a goldmine and I'm all for you doing anything you feel like doing with it and think it's really really wonderful to have! Whatever the methodological whatever, it's still scads more than we had, you know?

[dicta-contrion.livejournal.com]

On this I actually do have a hypothesis, which is that yes, there is more. But it's not actually based on anything! I don't keep track of that stuff at all. My empirics just get better and better. I just have a ~~feeling~~. But a feeling that I suspect will lead to more attentiveness to this going forward, so....idk.

Does offscreen sex count? I think if we know what happens offscreen, yes. If Harry wakes up the next morning with a sore arse or jaw, or tells poor traumatized Ron about it over breakfast, then I think we can reasonably take a guess.

And yes! Questions answered. And more (really good) questions raised, though I'm afraid it's late at night after a long day and I'm getting a bit punchy, so answers will have to wait until tomorrow. Will be back, though, and remain totally, totally intrigued. <3

[snowgall]

I think this is my favorite comment ever ♥♥♥

I sometimes feel like I'm the only person who gives a &%#@ when things were actually first written, and that I'm spending so much time on this endeavor for ridiculous reasons and no one will ever understand or appreciate why I care or how much time goes into it and I just want to hug you and cry over your comment and bask in the fact that somebody else gets it praise jesus and vishnu and the flying spaghetti monster

And it's awesome that you can consult with your own side-by-side statistician :) I also sometimes feel like a huge stats fraud. Like people will post comments like "oh this is cool I love stats" and I'm like, "dude, these are averages and counts and percentages. These hardly count as stats." If anybody asked me to do a chi-squared distribution analysis I'd duck and run for cover :)