Literary cognitive disonance or oversight?

I mentioned this earlier in another post, but a couple of authors have put together an analysis of authors and characters in young adult children's literature based on the winners of the American Library Association Best Fiction for Young Adults list which is published each year. They looked at the winners for 2011-2013 and posted their analysis here.

In many ways just another piece of flotsam and jetsam. But like a piece of corn stuck between the teeth, this one stayed with me, nagging. Their stated objective was to answer the question "How much diversity is included in the BFYA lists?" The first question I had about their analysis was why that question was pertinent. What was the agenda they were trying to advance? One can infer but they are not explicit.

But what was bothering me was a couple of omissions. They looked at representation by character and author of Race. That's a common one. Then they looked at character by orientation. Again, a common element that some people are especially concerned about. Then they looked at disability. Much more a marginal issue in terms of what people focus on but none-the-less a legitimate inquiry if one is focused on inclusiveness.

So what's missing? Well, nearly everyone with a concern about disparate representation in awards or mentions is concerned that women do not receive enough representation. No mention of gender at all in the analysis nor an explanation for omitting it. Seems odd to me.

They looked at identifying "Authors of Color" for the three years. Later when they looked at "Characters of Color" they had a very detailed breakout with all the common categories of African-Americans, Asians, Latino, but also the less commonly included Mixed Race and the rarely included Middle Eastern. OK, fair enough. But why did they report characters by these categories and not authors, instead lumping all non-Europeans into "People of Color"? And why did they focus only on the absolute measure of representation rather than the degree of over or under representation?

I think this is a case study of analytical cognitive dissonance. Typically in these types of analyses, the authors are attempting to advance three arguments simultaneously. It can't be said that is the case here, because the motivation for the analysis is not established. The common arguments are 1) the publishing industry is discriminatory against women and minorities, 2) that this is evidenced by disparate impact in terms of sales, awards, and other measures of recognition, 3) and that there is a harm arising from different groups' over or under representation.

As I have mentioned elsewhere, the measure of disparate impact we are actually most interested in, if we believe that there is actual discrimination occurring, is the comparison between the book buying population and demographic distribution of characters, authors, etc. When you do that, most of the apparent disparate impact disappears as there are material differences in book reading and book buying among different groups.

Setting that aside though, what are the answers that the authors omitted?

Over the three years I am counting 91 male authors (omitting three books that have male/female dual authorship) and 223 female authors among the total of 314 authors. That is a 29% male, 71% female split. In relative terms, given a 50:50 population ratio, males are 42% underrepresented and females 42% overrepresented.

The relative representation of characters is 20% underrepresentation for Asian Americans, 75% underrepresentation for African Americans, and 80% underrepresentation for Latinos.

LGBT characters are overrepresented by 71% (3.4% of main characters are identified as LGBT versus an commonly measured 1.5-2.0% of the population).

The over and underrepresentation of authors by race is trickier as the analysts did not include the numbers. But working from the pictures and other information in the post, it appears that 46% of the authors of color are Asian, 29% African American, 21% are Latino and perhaps 4% are Native American. Very rough numbers. Translating those figures to the population at large, that means that Asian authors are 229% overrepresented among AOC, African Americans are 17% underrepresented, Latinos 45% underrepresented and Native Americans right about at par.

I think the cognitive dissonance arises from these disparities. If one believes that publishing is in some way structurally biased against Peoples of Color and against Women, AND one believes that disparate representation is evidence of that bias AND one believes something has to be done about it, then you have to reconcile the irreconcilable. In other words, women are substantially overrepresented among the published authors (inconsistent with the theory of discrimination) and therefore one ought to advance initiatives to reduce that disparate representation.

Similarly with the authors of color issue. By aggregating into a hodge podge of Peoples of Color, you hide that Asian authors are dramatically overrepresented at the same time that others (AA and Latinos) are materially underrepresented.

I have argued elsewhere that these distributions are quite consistent with bibliocentricity (the degree to which people spend time reading books and spend money buying books). The critical challenge is to encourage increased bibliocentricity among all people and the effort invested in trying to identify active discrimination where no documentation exists is at best a diversion and at worst postpones the eventual achievement of equity that would likely be the consequence of uniform bibliocentricity.