yonmei

Hey - thanks for your comment on my statistical research on possible bias in the Hugos. Spurred by your comment that the Harry Potter and Vampire novels are actually Hugo eligible, I've gone ahead and used the Amazon sales data. I put the results up on your comment thread on feministsf, but for some reason it's taking forever for the moderator to actually add it to the thread, and I'd like you to see the data before the thread grows stale.

To put the summary up front, the Amazon sales data show that there is a bias in the Hugo nomination process. SF novels are nominated over 18 times more often than fantasy novels. This is significant because the data also show that while women make up just 24% of SF writers (supporting the earlier findings made using Gnod data) they make up 57% of fantasy writers. This raises the question of whether the bias is against fantasy novels, or against female writers. Looking deeper still, it turns out that men and women are nominated for Hugos in each sub-genre in proportion to their numbers in each subgenre, whether measured by writers or by books. Combining these two results suggests that the dearth of female Hugo nominations is due to a systemic bias against fantasy novels, combined with more females writing in fantasy, rather than to a systemic bias against females.

More interestingly, as a group females have a 59% chance of turning a nomination into a win, despite being just 24% of nominees.

The full post is below. Your comments are most welcome, you can reach me at paul@paulchafe.com

Cheers!

Paul



-----------------------------------------------
Harry Potter and Vampires are Hugo eligible? Hmmm – well, that certainly shows how much I track the Hugos.

However to be clear, I didn’t actually EXCLUDE HP or Vampire titles, I just didn’t use the methodology of counting Amazon.com titles because I erroneously thought it included non-eligible books. Instead I used a different methodology based on counting writers using Gnod. The statistics I quoted are valid for the sample I did use, and, whatever biases it may contain, an exclusionary bias is not one of them.

However, never one to pass up an opportunity to extend the database, I’ve re-run the numbers using Amazon’s top 100 sellers list.

Methodology

I used the top 100 selling SF titles on Amazon, and excluded anthologies (because I couldn’t necessarily get all the writers in them) and ambiguous names, and then repeated the experiment using the top 100 selling SF& fantasy titles on Amazon. It was WAY easier and faster to do this count than it was to do the same on Gnod.

Sources
http://www.amazon.com/gp/bestsellers/books/25/ref=pd_ts_pg_1

http://www.amazon.com/gp/bestsellers/books/16272/ref=pd_ts_b_nav

Bias

1) Excluding anthologies limits us (more or less) to the novel category, but that’s what I’m working with in the interests of faster research.

2) This is by no means a Hugo eligible sample. Mary Shelley is on there (and she probably should have a Hugo for Frankenstein) Anyone know if she’s got a retro?

3) These are bestsellers, which means they are popular. This is sometimes but not always the same as “good”. Presumably in the Hugos we’re looking to reward quality over quantity. However they are all included, to remove any personal bias I might have (actually I have lots) over what counts as “good”.

4) Amazon’s numbers shift constantly, and there may be artifacts in both their own methodology, and in the market. I know for awhile the top 100 was full of Star Wars titles, and now there aren’t any to be found.

5) As per usual, ambiguous names are discarded. I don’t think this is a bias, but it should be mentioned.

Data for writers in SF and SF&F genres found in Amazon top 100

SF titles (ie, no fantasy novels)

Count Percent
Male 68 76%
Female 21 24%

SF&F titles (including fantasy novels)

Count Percent
Male 38 43%
Female 50 57%

It should be noted that the SF&F list contains vampire writer Charlaine Harris no less than 20 times, and that vampire/horror fiction writers accounts for 38 of th

(Reply to this)

The writers who would have been shortlisted were:

Good grief.

(Reply to this)