holmesticemods (holmesticemods) wrote in holmestice,

Behind the scenes...Statistics!

Co-mod billiethepoet here to report some basic descriptive stats about the first nine rounds of Holmestice. colebaltblue did a lot of great work to turn all the header information participants have provided into a dataset so I could do some cleaning/recoding on it and whip up some charts and graphs. I am not a statistician but I do work in social science research, so I know just enough to be dangerous.

An Overview of Holmestice

So, we've completed nine rounds between Winter 2010 and Winter 2014. We've had a total of 552 fanworks submitted to the community by 280 unique authors/artists/podfic readers/etc for 291 unique recipients.

I am going to be peppering this write up with caveats, because those are the most interesting part of statistics to me: how the definitions we chose or are forced to work with shape the facts we accept. But feel free to skim through and just look at the pictures. :)

All of the data analysis for this was done in Excel because that's what I have and can work the fastest with for something like this. I used a Frequency Match formula to count the number of "unique" text entries in the 552 lines of data (each submitted fanwork is a line of data in the dataset). This means that if you created something with another participant, that counts as "unique" creator. For example, colebaltblue and I wrote a fic together and we've each contributed separately. So, using this method, we get counted as colebaltblue, billiethepoet, and "billiethepoet and colebaltblue". This artificially raises the number of unique creators but it was an uncommon occurrence so the effect should not be great. Also, if an individual submitted to the exchange under different LJ accounts, they are counted under each name. This could also cause some artificial inflation but there's no way for us to track that.

Over time, the number of submissions per round has stayed pretty stable:

The first round was by far the most popular then dropped pretty quickly to a more stable figure. There was a spike in the Winter 2012 round, which if I had to guess is probably due to a combination of the beginning of Elementary in September 2012 and the fact that Sherlock series 2 aired early in 2012. Lots of new canon that year.

Despite the fact that we encourage all types of fanworks, we are a fic heavy exchange:

Or, to look at it another way:

Fic makes up about 86% of the total submitted fanworks to Holmestice. A distant second is artwork with 12% of the total. Though we may not be diverse in terms of the type of fanwork being created, we are very diverse in other ways.

Sherlock Holmes Is Everywhere: Holmestice by Verse

We currently have 14 Holmesian incarnations represented in the exchange (including one original AU)!

You may notice that the total number of verses here adds up to more than our total of 552 fanworks. That's because some works were tagged as being in more than one Holmesian verse. We did not include any crossover verses in this analysis but that's something we could look at doing later.

Of those 14, only 5 have more than 3 fics total. So if we pull out "The Big 5" our percent distribution changes a bit:

(Not sure how I screwed up the data label but BBC 2010 is 76% of the fanworks in The Big 5.)

Another caveat is that these percentages are out of the fanworks included in The Big 5 verses only, not that overall total. The difference between the overall total and the The Big 5 total is small (565 total for all verses and 551 for The Big Five) so the percentage difference between the overall total and The Big 5 total would be pretty similar.

So let's just throw out BBC 2010. Let's ignore that behemoth for a minute:

In the "best of the rest" sort of image, we can see that ACD Canon is 57% of the rest of the fanworks in the exchange, with the Ritchie verse coming in second at 13%, then Elementary at 12%, and Granada at 8%. Everyone else is hovering around the 1-2% mark. Again, these percents are derived from the number of fanworks that do not tag BBC 2010 as a verse, not out of the overall total.

We are heavily dominated by BBC 2010 (please note that I did not say we were "dominated by Sherlock" because that would give you pervs ideas). But! Beyond that, we have a great variety of Holmesian verses represented. There's just not a ton of content represented across those verses, compared to Holmestice as a whole.

We're Not As Pervy As I Thought: Holmestice by Fanwork Rating

First, a note about ratings. We've had no consistent rating scale within or across rounds. Everyone just makes up what they want. There was a lot of similarity, particularly people using "movie style" rating systems early on and then moving into the AO3 rating scale. We decided to recode all of the ratings on the AO3 scale so we could do some comparisons and to make it easier to integrate data from later rounds into this type of analysis. I recoded the ratings on all 552 works into the 5 point AO3 scale: General Audiences (G), Teen and Up (T), Mature (M), Explicit (E), and Not Rated (NR). So, for example, a PG-13 rating became a T. They only tricky part to this is some fanwork creators like to "straddle" two ratings (like a PG-13/R rating on a work). In these cases, I coded up to the higher rating for safety.

We are very heavily slanted to T ratings! And G is our next biggest category! NR fics make up a very small category.

To look at that as percentages:

Anecdotally, I think we are increasing our instances of M and E fic as the rounds go on. There were very few in the first few rounds. That's an analysis for another time though.

What's Love Got To Do With It?: Holmestice by Pairing

In our 552 fanworks, we have 635 total pairing tags (including Gen as a tag). Of those 635, 57 are unique pairings! This can be couples, poly relationships/threesomes, gender swapped verisions of other pairings, etc. So, for example, Holmes/Watson is one pairing, while fem!Holmes/Watson, Holmes/fem!Watson, fem!Holmes/fem!Watson, Holmes/Watson/Lestrade, etc all stand on their own as unique relationships. These counts only include pairings that include at least one Holmesian character, so Hooper/Crieff counts but Doctor/Rose from a crossover fic would not. This is another place where the diversity of Holmestice really shines through.

Of the 57 unique relationship tags, 26 have only one mention. Because there's so much data here, I'm not going to talk about the majority of these pairings but know that if you're into a small ship, we probably have at least one fanwork for you.

There are only six pairing tags with more than six appearances in the exchange. We'll call these The Big Five + 1:

We are very heavily tilted toward Holmes/Watson (and that's just straight up Holmes/Watson flavor, without any of the possible groups, gender bend, or turning someone into a fawn options people have used before). Gen fic also makes a really strong showing with 158 total tags. Our third place pairing is much smaller but it is a canon het ship which I think is really sweet. :) I think it's also interesting that Holmes, Watson, and Lestrade each make three appearances in the top six pairings.

Since we are so heavily Holmes/Watson centric, especially if you consider all the possible permutations of H/W and not just the original flavor represented in the graph above, you may find yourself asking if a fic is Holmes/Watson:

Of the 635 relationship tags used, 52% of them are some form of Holmes/Watson. That includes Holmes/Watson and all the permutations of the tag based on gender bends, threesomes, etc. This is not done by fanwork itself so there is the opportunity for over reporting here. If a fic is tagged Holmes/Watson and Holmes/Lestrade/Watson, it was counted twice here since we were looking at the total occurrence of H/W in any form in the pairing tags. The takeaway is, even if you don't like Holmes/Watson, just under 50% of our output could still be something you're interested in.

Let's also take a look at what I'm calling "Relationship Composition" or couples vs threesomes vs gen fic. Note that our current poly ship tags are all threesomes. Somebody please feel free to write me a four or more some in the near future. :)

Gen fic makes up 25% of the total works in the exchange while 71% of fanworks include a couple and only 4% include a threesome.

So who are we writing about in these couples and threesomes though? We know it's heavily Holmes/Watson but let's look at a breakdown by gender composition of the pairings/groups.

It's no surprise that almost 60% of our fanworks contain a slash pairing. But 11% of the works contain a het pairing, 3% a femslash pairing, and 2% a threesome of mixed genders. Again, this count was done by tags so a fic could be tagged Morstan/Watson and Holmes/Watson. If that's the case, it shows up in both the het and slash categories. This is the most inclusive what to run that analysis but we could look at non-slash works that have no slash pairings as well.

I hope you've enjoyed some of these descriptive stats! This was my first stab at creating fandom stats and colebaltblue did an amazing amount of work on the front end getting the dataset together.

Feel free to leave questions in the comments or suggest any other analyses you might like to see. colebaltblue and I are considering doing this again on a larger scale.

Tags: !admin post, !behind the scenes

  • Post a new comment


    default userpic

    Your IP address will be recorded 

    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.