Analysis of Women's Publication Rates in Asimov's, Analog, and F&SF for the Years 1987 to 2001

John Borneman
12DEC2005

 

 

Table 1 - Average Publication Rates (1980 - 2001) by Magazine)

Table 2 - Average Publication Rates (1990 - 2001) by Magazine)

Chart 1 - Percent Women Published (1980 - 2001) by Magazine

Chart 2 - Percent Women Published  by sorted Total Magazine Averages and Total Year Averages

Table 3 - Analysis of Variance Results

 

 

Table 4 - Chi-Square Analysis of Asimov Magazine

 

 

Table 5 - Chi-Square Analysis of F&SF Magazine

 

 

Table 6 - Chi-Square Analysis of Analog Magazine

Introduction

This statistical study does not answer why. It attempts to shed light on some things that are true and some things that are not true. But it also raises some good questions for future study (see below). 

The data used information collected at the Broad Universe web site, and was inspired by Susan Linville's article. The data used was the number of published stories by women and men in Analog, Asimov's, and F&SF for the years 1980 to 2001. 

Any editors who wish to be statistically evaluated please let me know. no guarantees, but my statistical curiosity is pretty high.

Summary of Results:

  1. Asimov's and F&SF published a statistically significant higher percentage of women than Analog during this time period.
  2. Sometime around 1988, the percentage of women published took a statistically significant jump. This begs the question (not answered by this paper)…Why? What happened at this time?
  3. If we make the assumption that (a) women write stories that are as good as men's, and that (b) the percentage of women published should equal the percentage of women submitting stories to the magazines, it can be shown (given the errors inherent in this data) that; (a) Asimov's  published the same percentage of women as they receive in the slush pile, (b) that F&SF published more women than recieved in the slush pile, and (c) that Analog published less women than they received in the slush pile.
  4. A study of Andromeda Spaceways "blind-slush" process shows that there is not inherent prejudice against stories written by women with respect escaping the slush process. Also that the rate of women published in this magazine (30%) is far higher than the rate of women submitting stories to the slush pile (15%). 


Conclusions:


That the main issue with the apparent 30 to 40% publication rate of women is not due to prejudices against women writers but more due to the involvement of women in the submission process. It is admitted that data can be found which might show a significant difference between women's submission rate and publication rate, but that these are unique events and specific to certain markets and do not appear to be an indication of the industry as a whole.


Data Summary: See Tables 1 and 2



Graphical Results:

  • Plot by year of percent women published grouped by magazine. Is there a trend upwards? (See Chart 1)
  • Sorting the averages of percent women published by Magazine (left side) and by year (right side) Asimovs and F-SF seem to have greater percentages of women published during these years. Also, the jump of percent women published in the 1988 - 1990 timeframe is apparent. (See Chart 2)
  • Question not answered by this paper: Why the jump in percent women published during these years?



Statistics: 

  • An analysis of variance (ANOVA) confirms that there is a significant difference with respect to percent women published between years (confirms that the observed jump is real) and by magazine (Asimov and F&SF have higher percentages of women published than Analog. (See Table 3 )
  • A Chi-Squared analysis could be done to confirm that the percent women is not 50/50, but that is obvious from observation of the data. What is  interesting is to compare the submission rates of men and women (Expected Values) against the actual publication rates (Observed Values).
  • From Susan Linville's article she discovered through interviewing the major editors that Asimovs was receiving submissions from women at a rate of about 30%, the same for Analog, and F&SF rates were closer to 25%.
  • A Chi-Square analysis, which compares the actual (Observed) numbers of women published in each magazine over the most recent years of data available (1990 to 2001), with the "Expected" numbers that would be seen if the publishing percentage equaled the submission percentage revealed the following;

Chi-Square Test Set Up:

  • Null Hypothesis: Women and Men are published at their rate of submissions

  • Alternate Hypothesis: Women and Men are not published at their rate of submissions

  • Decision: if calculated Chi-Square (X2) is greater than the Chi-Square (X2) test statistic, then reject Ho and assume that the alternate hypothesis is true.

Results:

  • ASIMOVS:  X2 calculated (3.13) < X2 test (3.84), therefore Women and Men are published at a rate equal to their ratio of submissions. (see table )
  • F&SF:  X2 calculated (3.88) > X2 test (3.84), therefore Women and Men are NOT published at their ratio of submissions – they actually published more women than 'expected.' (see Table 5)
  • ANALOG:  X2 calculated (24.39) > X2 test (3.84), therefore Women and Men are NOT published at their ratio of submissions. They published less women than expected. (see Table 6)


NOTE: Remember that this data is from the years 1990 to 2001 and the observed upward trend shown above may have continued. 



Other Data:


As an interesting side point. Andromeda Spaceways Inflight Magazine (speculative fiction publisher out of Australia) has a unique system of slush reading that involves stripping the author identity away from the story before sending it to the slush readers. The make- up of these readers is about 55% women and 45% men, thereby negating any hypothetical selection preference based on gender- based story preference (i.e. men preferring "men" stories or women preferring "women" stories). Additionally, in order for a story to reach an editor's attention, any submission must be liked by at least 2 out of 3 slush readers 1. 


Looking through the current batch of stories, which were successfully were chosen for addition to the "slushpool," shows that stories written by men make up 80% of the stories, and those written by women only 15% (5% unknown sex). So stories by women authors make it out of ASIM's blind selection process at a much lower rate than men. I don't have access to the data that would tell me if these numbers match the actual men/women initial submission rate (but I suspect it does).


Also ASIM enjoys more submissions from the US than from Australia and also receives stories from Canada, Great Britain and other countries. Therefore the pool of authors submitting stories is globally diverse.

Ironically, going over the past nine issues of ASIM, the mix of "published" authors is 30% women and 70% men. 

Some Unanswered Questions:

  • Do men write different types of stories than women? This is a possible explanation for low publication rates by women in some magazines?
  • A collolary is, can one tell if a man wrote a story versus a woman? I have seen an online prose evaluator (can't find the link) that judges whether prose is written by a man or woman but it was not too good at it.
  • And, what happened in 1988 to 1990 that made women's publication ratios take a jump?
  • Any others?



1 The ASIM slush process is actually involves a ranking scheme but this description is accurate for these purposes.