The issue of representation of men and women in science fiction is one that has caused much discussion. There’s “The Count,” from Strange Horizons which looks at professional reviews of science fiction novels by gender, and a similar initiative by Lady Business for blog reviews. A common query when it comes to this issue of representation is the makeup of submissions. Are the differences we see simply reflections of the proportion of submissions received?

There has been some investigation of this question before, and by widening the data-gathering net, I hope to allow a more nuanced and informed discussion of the issue. I have also attempted to perform separate analyses of submissions as a whole, as well as science fiction submissions specifically.

In the previous article I looked at the gender ratios of published stories of all genres and published science fiction stories in seventeen SFWA short fiction markets. While looking at published fiction alone does give us some important information about the gender representation of authors both overall and between different markets, it’s not the end of things. In addition to gathering data on published stories, I asked editors for details of the source of those published stories.

With some editors receiving hundreds of submissions every week, the effort and time to categorize even a sample of those submissions for this study was substantial, and I am extremely grateful for that assistance.

Publications from Slush

Some markets draw all of their published stories from open submissions, while others draw from a combination of slush submissions, reprints, and stories solicited from authors. For the second group, it is useful to compare the published stories with those that came from slush, to see if this results in significantly different gender ratios. After all, it could be the case that reprints or solicited stories show a different kind of gender ratio than the stories that just come from slush.

Those markets which drew published stories from outside the slush pile during the period studied were: Apex, Clarkesworld, Daily Science Fiction, Escape Pod, F&SF, F&SF Special Issue, Lightspeed, Strange Horizons, and Tor.com.

Table 1: Reprints and Solicited Stories Status of Markets

Taking this subset of markets and graphing Publications from Slush vs Total Publications, we can see that reprints and solicited stories make up a large proportion of published fiction in some markets. When looking at the percentage share of each gender, there are some slight differences when comparing Publications from Slush with Total Publications for both all stories and science fiction stories.

However, overall, the differences in ratios between All Published Stories and Stories from Slush are not significant.

Submissions: Samples and Variation

As mentioned in the previous article, please note that this is a study of apparent gender. Some editors used Google to obtain public bios of the authors for a more accurate categorization of author gender, and some were already aware of the gender identity of some authors, but in most cases the categorization by gender was based on first name. This inherently makes the gender split less accurate than it was in the previous article.

Additionally, when dealing with overall publications and comparing them to publications drawn from slush, I could look at the full data-set for the time period. However, with some publications receiving thousands of submissions per year, many markets were understandably unable to provide full breakdowns for a year’s worth of submissions. As such, I am dealing with samples from varying time periods. This means that the analysis will not be quite as comprehensive or accurate as would be ideal.

Drawing conclusions about the yearly submissions from a smaller sample may lead to error. For example, Escape Pod provided data for the months of February 2014, April 2014, and May 2014. We can see a variation of 4.01% between the highest and lowest months in terms of gender ratios.

Table 2: Escape Pod Submissions by Month

Looking at a separate twelve month period of Clarkesworld Magazine’s submissions, we can see additional evidence of variation in gender ratios. This chart looks at both total submissions (including science fiction, fantasy, horror, science fiction/fantasy and science fiction/horror) and science fiction subs only (including science fiction, science fiction/fantasy and science fiction/horror.)

While the overall breakdown for the year was 71.06% men and 28.94% women for all submissions, the highest month for representation of men came in at 73.19%, and the lowest at 69.62%. This gives us a 3.57% difference between the months with the highest and lowest representation of men. So, we can see that a one-month sample would not have accurately reflected the yearly ratios.

Table 3: Clarkesworld Submissions Gender Ratio Variation

We also see variation in overall volume of submissions, as detailed below. (In all cases but the highest volume of total subs by women, the highs/lows in volume by gender were also found in the same month as for overall submissions.) Again, we can see that while the average (mean) month had 776 submissions, with 551 by men and 224 by women, one month had 933 submissions total, while one had only 649.

Table 4: Clarkesworld Submissions Volume Variation

Given these variations, it is likely that the submission samples are not giving a completely accurate representation of the overall yearly ratios, so any kind of specific ranking of markets based on the ratios given here would not be a good idea. That said, we can likely still draw tentative generalized conclusions.

Overall, I have data for fourteen of the seventeen markets when it comes to submissions for all genres, and twelve of the seventeen markets for science fiction submissions.

Table 5: Submissions Data Provided By Markets

Submissions Volume

In addition to looking at the proportions of submissions by gender, it is also interesting to look at the number of submissions received by each market. There is a wide disparity between the volume of submissions received by each market. In order to give an indication of this, markets that gave only a sample of submissions were brought to a one year volume for comparison. This is only a rough comparison; in addition to the variability between different time periods mentioned earlier, some of these markets may be closed to submissions at various points.

The deficiencies of this method of estimation can be seen by comparing the actual total submissions with the estimate from the sample, in the markets where we have both values: Flash Fiction Online, IGMS, and Bull Spec. This has some impact on certain statistical tests (e.g. chi-square) but does not necessarily invalidate most of the results found here.

Submissions Ratios

Overall, we see a much smaller degree of variance between markets with Submissions than we saw with Publications.

All Submissions

Range - Submissions by Men: 81.31% to 54.59% = 26.72%
Range - Submissions by Women: 18.69% to 45.41% = 26.72%

The mean ratios are 67.94% men, 29.63% women, 0.02% non-binary and 2.43% unknown.

The median ratios are 68.40% men, 29.55% women, 0% non-binary and 0% unknown.

Science Fiction Submissions

Range - Submissions by Men: 81.31% to 58.67% = 22.64%
Range - Submissions by Women: 18.69% to 41.27% = 22.58%

The mean ratios are 70.92% men, 27.52% women, and 1.61% unknown.

The median ratios are 72.65% men, 25.89% women, and 0% unknown.

Note: For ease of comparison between All Subs and Science Fiction Subs (where possible), the source of the data is indicated in the label for each market.

No market received more submissions from women than from men.

As in the previous article, I tested to see if these differences were significant, and found:

The differences in submissions ratios between markets are significant, both for all submissions and science fiction submissions specifically.

Science Fiction Submissions by Genre of Market

I have no evidence to say that women make up a smaller proportion of submissions for science fiction only markets than for mixed-genre markets.

Mixed-Genre Publications: Gender Split of Total Subs vs. Gender Split of Science Fiction Subs

I found a moderate-strong positive correlation, so there is a tendency for mixed genre-publications who receive a high proportion of total submissions from one gender to also receive a high proportion of science fiction submissions from that gender.

Submissions Ratios by Gender of Senior Editorial Team

I found no significant relationships between gender of senior editorial team and gender ratios of submissions either for all stories or for science fiction stories.

Story Submission Split by Average Age of Senior Editorial Team

I found no significant relationships between age of senior editorial team and gender ratios of submissions either for all stories or for science fiction stories.

So, What Does This All Mean?

There are several interesting results from looking at the Submissions data. While there is significant variation between markets, the range is much smaller than we saw in Publications. In addition, while a significant number of markets published more women than men in the Publications data, no market received more submissions from women than from men, either overall or for science fiction. In mixed-genre markets, we see that a relatively high proportion of total submissions from one gender is correlated with a relatively high proportion of science fiction submissions from that gender.

The volume of submissions received shows dramatic variation. While these figures were estimated in some cases, it is fair to say that yearly submission volumes between markets range from a few hundred up to a few thousand. As such, the idea that gender ratios in submissions are due to the fact that women simply don’t write science fiction is patently untrue. While this might be the case for high-volume markets, the number of science fiction stories written by women in total is far greater than the number received by many markets. For example, if approximately 30% of the science fiction submissions by women received by Strange Horizons were instead submitted to Nature, we would see an equal gender representation in Nature’s submissions. (This is assuming that all stories submitted to Nature were also submitted to Strange Horizons. If we assume no overlap, the number is around 20%. The actual number is likely somewhere in the middle. Please note that I am aware that a suitable submissions for one market is not necessarily a good fit for another. This thought experiment serves as an illustration of the comparative volume of submissions and the effect that increased or diverted submissions from a small percentage of the women authors in the field could have on some markets.)

However, there is no obvious smoking gun which shows us why submissions ratios vary between markets. I found no significant correlations between gender ratios of submissions and any of the categories that we tested. Based on our data, genre of market, age of senior editorial team, and gender of senior editorial team, are not factors that are related to the ratio of submissions by men and women.

A Note on Comparing Submissions with Published Stories

This all begs the question—is there a relationship between submissions ratios and publications ratios? In the next installment, I will be looking at what further analysis can be carried out regarding this data. In the meantime, I would like to warn readers about the problems with attempting to carry out a simple analysis of the Submissions Ratios from samples, vs Publications From Slush for a whole year.

Conveniently, we have a good example of the pitfalls of this approach. Four markets provided data on the Stories Accepted from Slush during the time period of submissions data provided, which we can compare to Stories Published from Slush in the time period given in the previous article:  Clarkesworld, Bull Spec, Flash Fiction Online, and Lightspeed.

Looking at the differences between what was accepted from that specific slush sample, and what was published in a full year, we can see that a simple analysis of Submissions Ratios for the samples given vs Publications from Slush for an entire year would not give an accurate picture.

Table 6: Acceptances from Submissions Sample Time Period vs Yearly Publications

Next Time on Adventures in Statistics: Our hero wraps things up with a pretty pink bow! Will she be able to avoid overdosing on graphs?

Share this page on: