From: Charles Bayless <charles.bayless_at_ttmd.com>
Date: Wed, 27 Feb 2013 10:23:43 -0500

I keep asking for evidence regarding the measured prevalence of diversity and conformity match (where critical identity variables match between the reader, the author and the protagonist) in children's literature and whether there is any measurable impact from presence or absence of conformity literature. I can't answer the second question but I decided to take a back-of-the-envelope crack at the first. In doing so, I find that there are actually two elements to the question - 1) to what extent are readers willing to read beyond their own culture and key attributes and 2) to what extent are Group X members un- or underrepresented in children's literature. We often speak of diversity and multiculturalism as if they are synonymous with Group X underrepresentation. It appears that they are not synonymous.

I took about four hours to do a quick and dirty measurement of conformity literature, multiculturalism, and diversity to try and wrap some numbers around an ill-defined topic. What specifically does it mean when we call for greater diversity, more multiculturalism, and more individuals with whom children can identify in children's literature? I preface this exercise with the firm disclaimer that these are all good faith but necessarily ballpark numbers.

Just the exercise of figuring out what to measure and how, was illuminating and clarifying. I have defined diversity as reading about people who do not share your primary Group traits. I have defined multiculturalism as reading about protagonists who do not share your own culture. My conclusion from the exercise is that 1) children's literature is already incredibly diverse so trying to increase diversity is probably not going to make much difference, 2) children's literature already has a high degree of multiculturalism ranging from 35% (by author) to 60% (by protagonist) so further efforts in this arena are also likely not to make much difference, and 3) despite the existing diversity and multiculturalism, the odds of a particular child, recognizing themselves in the identity of authors and protagonists is low to start with and becomes exponentially lower, the more attributes you add to the Group X definition. Finally, 4) the economics of small numbers wreaks yet further havoc upon compound Group X factors
. If you are a YA GLBT Native American, the very rough estimate of market size would look something like: 5 Million Native Americans X 1.5% (rough percentage of GLBT population) X 44% (Number of households with children) X 50% (percentage of the population who read recreationally) X 10% (chance of a person buying a particular book in which they are interested - made this number up and assigned a high side value, all the other numbers are sourced) approximately 1,650 book sales if everyone who is a likely and interested reader with attributes in the target population buys a copy. Obviously that is just an order of magnitude estimate and there will be sales beyond the YA NA GLBT population but it does give a sense of the financial risk involved. With that small a market, there is no room for errors in cost.

With my self-imposed limit of 4 hours, I was not able to look at the top 100 books published in 2012 and do this analysis. Instead, what I did was two-fold.

First I looked at ten books to which I was emotionally attached as a child/YA or which materially changed a point of view that I had. In other words, books that were in some way significant and with which I connected. The books were Mike Mulligan and His Steam Shovel, The Little Engine That Could, Jonathan Livingston Seagull, The Teachings of Don Juan, Kon-Tiki, Little House on the Prairie, Pippi Longstocking, Sherlock Holmes, William Flinders Petrie (I had a long standing fascination with Egyptology), and A Night Before Christmas illustrated poem. I looked at each of these ten and assessed my personal degree of conformity on three attributes and in regard to two variables. Specifically, did my Race, Culture and Gender (RCG attributes) match that of either Protagonists or Authors (the variables)? If these were important books in my childhood then one might expect, if conformity is a real and important issue, that I would have a high conformity measure (with Protagonist and Author) for these important books. In fact, my RCG did not match that of any of the protagonists and only two of the authors, indicating that conformity has only a limited predictive power in terms of which books are going to be relevant/important to a child (with the caveat of course that this is based on a sample of one).

Taking a slightly different perspective, I looked at the RCG of the ten protagonists and asked what degree of overlap there was with my own RCG? Example: if the sample consisted of ten books of German literature by women authors, I would have an Authorial RCG overlap of 33% (10 white race, 0 from same culture and 0 from same gender, 10/30 33%). My Protagonist RCG overlap for the actual ten books was 40% and my Authorial RCG was 63%. So for these ten critical books, the scores were

RCG Protagonist Match - 0%

RCG Protagonist Overlap - 40%

RCG Author Match - 20%

RCG Author Overlap - 63%

Total Author and Protagonist conformity - 0%

Obviously this is subject to criticism but it at least begins to put some boundaries around pure speculation. Ideally what you would do is to look at more variables than just RCG. In terms of identity, likely ones ought to include Class, Religion, Familial Status, and Orientation with plenty of other candidates (Health/Morbidity, Income Quintile, etc.). You would look at a much larger population of books (in the hundreds). You would look at the current population of books being read by children (not necessarily the same as those being published). You would use standard definitions to provide rigor. You would include not only protagonists but all characters within the book with whom children might identify.

This sort of analysis was way beyond the 4 hour constraint. I took an alternative approach. I maintain a database of some 20,000 titles selected by durability and popularity as determined by frequency of awards received, mentions on library lists, mentions by the reading public, etc. It is not super rigorous but it is pretty good. You would recognize all the books (examples among the top titles: Charlotte's Web, Where the Wild Things Are, Goodnight Moon, Mr. Popper's Penquins, Curious George, A Wrinkle in Time, Madeline, Make Way For Ducklings, The Secret Garden, The Little House). I took this list and rank sorted by greatest popularity and then selected the top 100 books that I recognized from my childhood from among those most popular. By using books from my own childhood, I was able to identify the RCG elements without any research, i.e. quickly. By using books from my childhood I bias the sample by those childhood circumstances (American but raised abroad in several different countries where books were o ften less than readily available). I also bias the sample by age, i.e. no contemporary books. All that said, it still moves things forward beyond simply speculating.

Another drawback of this database is that, by tracking popularity, it is overwhelmingly fiction and tends to omit non-fiction books which for many children constitute a material percentage of their reading. Probably 60% of my childhood reading was non-fiction and some of the influential books that are also popular (such as A Night to Remember by Walter Lord; The Guns of August by Barbara Tuchman; Tutankhamun by Desroches Noblecourt; Gods, Graves, and Scholars by C.W. Ceram, etc.) simply don't lend themselves to RCG classification.

The importance of definitions became critically clear when conducting this RCG analysis on 100 books instead of ten. What do you do when there are multiple authors? How do you assign RCG to fantasy characters (What gender is the Velveteen Rabbit? What culture is the Little Engine? ) How do you assign authorial culture to authors who were born and raised in one country but later migrated and took citizenship in another? Is there a time limit on culture (i.e. Is it accurate to say that you share the same culture with an author from your own country but who was writing in the 1860s? Heck, are children of today of the same culture as 1950?) How do you deal with identity definitions (should Sydney Taylor be considered the same culture because she is American or a different culture because she is writing within the Jewish tradition/experience? Is Anne of Green Gables a different culture because it is Canadian or the same because it is North American?) And so on. And that is only for three simple variables of Race,
 Culture and Gender. I made the best judgments possible within the 4 hours.

The results then from this RCG analysis of Protagonist and Author for the top 100 American children's book with which I was personally familiar are:

RCG Protagonist Match - 18%

RCG Protagonist Overlap - 50%

RCG Author Match - 32%

RCG Author Overlap - 73%

Total Author and Protagonist conformity - 15%

So for a sample of 100 books of the roughly 800-1000 that I read as a child, I had conformity alignment (complete shared RCG attributes with both author and protagonist) with 15%. Material, but not convincingly demonstrative that that is a critical issue in reading. So for an average child then, that implies that they are likely to see themselves in the author AND protagonist attributes only 15% of the time.

Even if the conformity index is very low, what about diversity and multiculturalism? Since I am in the second largest demographic group (White, American Culture, Male), and assuming that the largest group's numbers (W,AC,F) are comparable, what this indicates is that there is a very high degree of diversity and multiculturalism already. In other words, if you are not from the numerically dominant group, your alignment numbers will inherently be even lower, i.e. even more diverse. Some specific numbers for those 100 books:

Authorship Race: 99% white, 1% other

Authorship Culture: 67% American, 33% other

Authorship Gender: 54% male, 46% female

Protagonist Race: 59% white, 41% multiple other

Protagonist Culture: 39% American, 61% other

Protagonist Gender: 53% male, 47% female

The authorship race stands out and is reflective of books written more than forty years ago. I suspect that the authorial race diversity would be far higher today, particularly if you look at number of titles published rather than just the number of books sold. In fact, I suspect that all the diversity and multicultural numbers would be higher today (but leave that for someone else to calculate).

So if kids are already reading highly diverse and multicultural books in terms of authors and protagonists, and the reading portfolio is already getting more diverse and multicultural, what benefit will be derived from adding another few percentage points? I think what this indicates is that multiculturalism and diversity are not the real issues. Children are willing, and able, and actually do read books about characters who are not like them and by authors who are not like them.

What the analysis also indicates is that even in an environment of high multiculturalism and diversity, you can still have reading individuals who fail to see themselves reflected in the body of books available. If I am one of the 1.5% of Americans with a Native American background (or 1.5% GLBT, or 6% Jewish, or 20% bottom quintile income, or 13% African-American, etc.), I am unlikely to find all that many conforming authors and protagonists among the broad portfolio of books that others are reading. If the conformity isn't there for the majority group, it will be even more dramatically absent for smaller groups.

So the push by some to encourage authors to keep in mind the rich diversity of the US when they people the characters in their books makes sense but probably doesn't get us very far.

If the reading public is already open and accustomed to non-conforming, multicultural, diverse reading AND there is effectively a near limitless supply of new titles AND there are low barriers to writing and publishing for small markets AND there is high publishing risk (evidenced by a high failure rate among new books with low sales and no profit), then the solution to increasing the opportunity where people can see themselves based on whichever attributes are most pertinent at the time, seems to reside in getting people in that Group X to read more and buy more of these books. Simply producing more won't solve the problem unless there is demand and unless there is a means of connecting that demand to the supply. And I recognize that this approach also has all sorts of issues (who belongs to Group X, is their depiction authentic, what to do when the representation is statistically accurate but might fuel a negative stereotype, etc.) People are willing, able and do read books that are both diverse and multic ultural. In an increasingly heterogeneous population with multiple attributes by which to identify a Group as X, and where Group X may be a compound definition of attributes, the potential market of Group X can quickly become too small for commercial purposes with too little prospect for upside potential. Under these circumstances, I keep coming back to the five actions that I think are at the heart of this issue (regardless of which Group X is the focus).

1) Increase demand for books in general and the habit of reading in particular within their constituency (as well as at large)

2) Increase quality of books (broadly defined but especially in terms of editorial review)

3) Increase the cultural and societal value attached to enthusiastic reading

4) Improve or create better market making mechanisms for matching supply with demand

5) Improve forecasting competency to improve the yield of profitable books to the total number of published books

Jeffrey Brenzel (Dean of Undergraduate Admissions at Yale University and philosopher) defines a classic as 1) Addresses universal and permanent human concerns, 2) Is a game-changer, 3) Influences other great works, 4) Is respected by experts, and 5) Challenges as it rewards. In the absence of any standard definition, that looks like one that most people might agree with most elements.

Accepting that, it does then pose an irony. The more you target smaller groups, the less able you are to meet those five requirements of a classic and therefore the less likely you are to have a winner that bursts the boundaries of group definitions.

I think we have heard several comments or observations in this discussion to the effect that we ought to be seeking the universal within the framework of the particular; that that is the only way to get the numbers and the broader popularity that would allow a sustainable success. That does seem to be the only way: to cut the Gordian Knot of limited demand arising from particularity with the sword of universalism.

Charles
Received on Wed 27 Feb 2013 10:23:43 AM CST