From: Charles Bayless <charles.bayless_at_ttmd.com>
Date: Wed, 27 Feb 2013 10:23:06 -0500

I have been thinking about our discussion from the beginning of the month. I wanted to return to it before we transition to the next set of discussions for March. Broadly the set of issues raised have been about diversity, multiculturalism, identity, representation in literature, and stereotypes. My earlier comments were basically an observation that we talk about these topics a lot but never go anywhere with them and that I thought part of the reason was that we do not have settled definitions, we talk at cross-purposes and that there is an over-reliance on assumptions and opinions and an absence of data to support many of those assumptions.

The argument that has been made in one form or another is that some children are underrepresented in children's literature and that that underrepresentation does cause measurable harm to them. There is a corollary argument that only authors who share the Group X attributes can authentically represent those Group X children in literature. The root cause is ascribed to conscious or unconscious bias or discrimination within the publishing industry. There is usually a set of solutions which encompass exhortations to publishers to publish more Group X books and the need for more or better publicized Group X awards.

My initial response was two-fold - 1) Do we know if there are real, measured disparate representations of different Group Xs (where Group X is any defined group with a shared attribute or portfolio of attributes) and to what extent are there disparate representations, i.e. which groups are over and which groups are underrepresented and by how much? 2) Is there any evidence that this causes a problem for children? Can it be demonstrated that an absence of characters with matching Identity traits does some measurable harm to children with those traits? Based on the discussions and further research, I have a third question. 3) Is there any evidence that the identity attributes which we as adults think are critical (Race, Culture, Gender, etc.) are the actual identity attributes by which children identify themselves?

The answers to these questions seem to be 1) No, we do not have any measure of which groups are over or underrepresented or by how much. We know there is disparate representation but no one has measured it. 2) No, there is no evidence that children are measurably harmed or benefitted by degree of self-representation in children's literature, and 3) No, there is no evidence regarding which attributes are most pertinent to a child's own self-definition.

To elaborate briefly on points two and three. The marked life success performance of disadvantaged immigrant groups in the past fifty years who have approached or exceeded native born life success measures (education attainment, income, morbidity, familial stability, etc.) is one significant counterfactual that calls into question the contribution of self-identity to life success. Nigerians, Asian Indians, Pakistanis, Koreans, Haitians, Dominicans, etc. are all recently arrived (within two generations) with few or no migrant antecedents. Consequently their first, second, and third generation off-spring have grown up with little or no representation in children's literature and yet have demonstrated high levels of accomplishment. This does not completely refute the assumption that self-representation might be important but it does call it into question.

With regard to the third point, the work of Henrich, Heine, and Norenzayan, The Weirdest people in the World?, calls into question fundamental psychological precepts and issues of self-identification. Their paper is here, http://tinyurl.com/4jvmv36. A narrative summary and background is here, We Aren't the World, http://tinyurl.com/b4a9fqg.

I have expressed concern about the absence of a measurement of the problem, doubt about the likelihood of measured negative consequences, skepticism about both the proposed root causes and the likely efficacy of the proposed solutions. If I am right, then the existing line of problem definition and problem solution leads to unproductive actions and failure to achieve the desired goal. It is inadequate to simply express doubt, it is incumbent to show evidence and probability that the argument is wrong, and best to propose alternative pathways towards success. To that end, I have a series of posts attempting to lay out the argument why the traditional position is wrong and in pursuing that analysis, I seek to identify what alternative actions might be more productive.

Typically in problem solving, you first define the nature and magnitude of the problem, then demonstrate that it does exist as a problem, then show that there is an undesirable consequence of the problem, next determine what are the real root causes of the problem and then formulate a series of actions that can be demonstrated to ameliorate the problem without causing off-setting costs and unintended consequences.

For purposes of this discussion, I have taken out the issue of stereotypes as too hydra-headed for an already complex discussion. There is a rich discussion to be had about the definition of a stereotype, the validity and utility of stereotypes, the relative problems and benefits of positive versus negative stereotypes, the distinction between invidious negative stereotypes and stereotypes that have a statistical basis, the use of stereotypes as literary shorthand, etc.

I am focusing here on identity, diversity and multiculturalism. I am defining the problem statement along the lines of "Children deserve a rich reading experience that encompasses the world as they know it and the world that is beyond their experience. The reading experience should leverage their interests and capabilities and should be sufficiently rewarding that they wish to and do continue to read. Too many children, for many reasons, do not enjoy reading, do not read much or do no elective reading, thereby constraining their world and their opportunities to thrive in that world."

With that as the problem statement, I am simply addressing the proposed hypothesis that the identities of children in multiple Group Xs are underrepresented and that this underrepresentation has a measurable negative impact and is therefore a material cause of the problem statement. If that hypothesis is correct, then it shapes our actions in one direction. If it is wrong, then we head in a different direction. The goal of providing a better, richer reading experience to children remains the same.

I acknowledge that we live in a highly heterogeneous country in a rapidly interconnecting world; exogenous circumstances which affect this analysis. I also observe that there are numerous circumstances which make the US an outlier among OECD countries - particularly its federalism, pluralism, individualism, heterogeneity, religiosity, etc. While I believe we all share the view that the ability to read capaciously, broadly and critically is a cornerstone for life success, that is a relatively difficult argument to make on an evidentiary basis. Principally, the challenge is one of causation. We know that enthusiastic reading is associated with such positive life outcomes as education attainment, familial stability, longevity, health, income, wealth accumulation, etc. The challenge is to establish causation. Does reading cause these things or is there a more fundamental cause that leads to reading along with all these other benefits. I am setting that argument aside as well for this discussion and simply take a s axiomatic that enthusiastic reading is beneficial.

I acknowledge that making an argument in the social sciences field is fraught with issues of politics and ideology as well as inadequacy of evidence. The error rate, non-replication, and retraction rate for papers in the hard sciences exceeds 50%. The error rate in the field of the social sciences is much higher with the further complication of numerous incidents of fraudulent research. Social sciences is a field with a scarcity of solid, commonly shared facts which makes it difficult to reach agreement on critical issues and where beliefs often take the forefront over evidence. It is for that reason that I originally attached the bibliography regarding group decision-making. These issues in the social sciences are not new or unknown; there are ways to be aware of the pitfalls and to address them. My inquiries are not an effort to insult anyone's belief system. All I am trying to get to is 1) What do we know?, 2) How do we know it?, and 3) Can we use that knowledge productively to benefit the reading experie nce of children? In a world of scarce and tightening resources, and high rates of change and uncertainty, it is incumbent for all decisions to be well reasoned and fact based.

I also acknowledge that there is almost certainly disparate reading outcomes by any of a variety of identity attributes. It is true in all other fields of endeavor, whether academic, cultural, scientific, economic, sports, etc. that there are disparate outcomes by identity attributes. It is unlikely that every other field would have disparate distributions and not reading. My argument is not whether there is disparate impact but rather what is the measured variation. Which groups are overrepresented and which ones are underrepresented and by how much.

More fundamentally my question is whether such a disparate impact has any measurable negative consequences as is alleged. Nobody has proffered any evidence that there is any measured impact and there are plenty of counterfactuals to support the hypothesis that identity reading is not consequential in life outcomes.

What follows is a series of posts about identity, diversity and multiculturalism in children's literature. The summary below allows you to skip the next set of posts. It is based on the prior couple of posts as well as the ones about to follow: Reading Diversity Analysis; Mathematical Necessity of Diversity; Market Structure; Disparate Impact; and Anecdotal explanations unless you are interested in the details. I especially encourage consider reading the following posts, Reading Diversity Analysis and Mathematical Necessity of Diversity which sheds some numerical light on identity in children's literature.

Conclusions

Readers are catholic in their tastes and read with great diversity and enthusiasm across cultures

Even though they read diversely and cross-culturally, there are still variations in literature adoption between different Group Xs

The probability of any child in any Group X identifying themselves based on attributes (such as RCG) is very low, even if the child is in the majority element of any of those attributes.

The production of new books is growing faster than the base of readers

The supply is effectively infinite and the demand is limited

Reading demand is a pareto distribution with a small percent of the reading population driving a disproportionate amount of the demand. If those enthusiastic readers are concentrated among specific Group Xs, they likely create distortions in the nature of the demand market for books

The size of the potential market for many compound Group Xs (i.e. multiple attributes) makes commercial success less likely given existing production costs and especially the expense of matching readers to specialty literature

Statistically improbable as any combination of Group X attributes might be, there is still the possibility that there will be a book with those attributes that achieves pertinence, perceived excellence and/or popularity.

There is no predictive model or other algorithm which usefully or accurately forecasts the likely critical reception, popularity or durability of new titles, particularly those by new authors

Experts within the field routinely judge particular books as excellent but which do not engage the reading public over the long term

Experts within the field routinely overlook the bulk of new titles that do achieve durability

Group X authors can sometimes have a proclivity for writing about pivotal event X, potentially fostering tragedy fatigue and indictment fatigue which reduce demand

Long tail competition reduces the odds of success for each new offering, reducing publisher interest

Markets are smaller than are commonly estimated

Demand is lower than might be anticipated

The mean time of research is greater than the mean time of decisions that have to be made

The mean time of implementing decisions is greater than the mean time of change

Research usually plays catch up to new developments and follows change, not lead it

Incentives work: make it easier, cheaper and more rewarding for that of which you want more. Make it harder, more expensive, and less rewarding for that of which you want less

New titles and new authors represent the greatest risk for the financial well-being of publishers (with the possibility of great reward).

Opinion

The human system is complex, chaotic, non-linear, with multiple internal feedback mechanisms and is inherently unpredictable. As such, there is no single root variable that can be addressed to achieve a desired outcome. Every problem has multiple root causes, unperceived connections with other issues, involve undesirable trade-offs and frequently entail unanticipated consequences. To achieve a desirable outcome, set a goal, set parameters for action and execute adaptively. (Metaphor alert) There is no low-hanging fruit, there are no silver bullets, there is no easy solution, and there is no sure thing. To paraphrase Hartley, The future will be a different country and they will do things differently there.

Possible Solutions

In suggested order of priority.

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

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

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

4) Identify, leverage and network Group X infrastructure to identify and amplify successful Group X books and to increase volume of sales, especially cross-over sales beyond the boundaries of Group X

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

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

7) Encourage Group X authors to range beyond Event X

8) Encourage Group X authors to write universal stories with the attributes of Group X

9) Shift the critical focus to that of the MLK approach of identifying by character attributes rather than demographic attributes

10) Find a mechanism to bring the common interests of the Academy, Teachers, Librarians, Publishers, Authors and Booksellers into greater alignment behind concerted actions constructive of more enthusiastic critical reading.

Research topics

1) What books do children actually read, how many, and at what age?

2) Why do they choose those particular books, and under what circumstances do they read them?

3) How do they respond to those books that they do read and with which characters do they identify, to what extent and why?

4) What are the proportionalities of the traditional identity markers 1) in the cannon of children's literature, 2) in new titles published each year, 3) and in the books that children actually read each year?

5) Is there any causative relationship between identity representation in children's literature and in life outcomes (education attainment, employment, familial stability, income, wealth accumulation, morbidity and mortality, etc.)?

6) What are the measurable, predictive variables associated with low levels or absence of reading?

7) What are the variables that have predictive capability between new titles and popularity (volume of demand)?

8) What are the variables that have predictive capability between new titles and durability (on-going demand over time)?

9) What are the attributes by which children self-identify and are those attributes predictive of the characters with whom they identify in books?

10) How often do those self-identity attributes change, and do they change in any predictable fashion (for example by age)?

Group Decision Making Bibliography

Bibliography of books related to decision-making in diverse and uncertain environments.

What The Numbers Say by Derrick Niederman and David Boyum - Using factual analysis as the basis for insight and decision-making

Statistical Analysis for Decision Making by Morris Hamburg - Using rigorous statistical tools in individual and group decision-making.

The Signal and the Noise: Why So Many Predictions Fail - But Some Don't by Nate Silver - How do we distinguish new knowledge that is real from simple noise and how do we become better at making accurate predictions?

The Half-Life of Facts: Why Everything We Know Has An Expiration Date by Samuel Arbesman - How new knowledge evolves and why experts are so often wrong.

Risk: The Science and Politics of Fear by Dan Gardner - Perceived risk versus quantified risk and their respective roles in individual and group decision-making.

Against the Gods: The Remarkable Story of Risk by Peter L. Bernstein - Using risk in forecasting and decision-making.

Reckoning With Risk by Gerd Gigerenzer - The role of risk and uncertainty in decision-making.

The Collapse of Complex Societies by Joseph A. Tainter - Decision-making in the context of complexity, risk and sustainability.

The Progress Paradox: How Life Gets Better While People Feel Worse by Gregg Easterbrook - Well-being, progress and measurement.

Nonzero: The Logic of Human Destiny by Robert Wright - Complex cooperative group decision-making.

Linked: How Everything Is Connected to Everything Else and What It Means for Business, Science, and Everyday Life by Albert-Laszlo Barabasi - Complex cooperative group decision-making.

Thinking, Fast and Slow by Daniel Kahneman - The intersection of individual and organizational decision-making.

Trust: The Social Virtues and the Creation of Prosperity by Francis Fukuyama - The role of trust in group decision-making.

It Ain't Necessarily So by David Murray, Joel Schwartz, and S. Robert Lichter - How common knowledge is often wrong and undermines effective decision-making.

Being Wrong: Adventures In the Margin of Error by Kathryn Schulz - Explores overconfidence in undermining effective decision-making and the valuable role of acknowledging error as an impetus to improvement.

Why Most Things Fail: Evolution, Extinction and Economics by Paul Ormerod - The critical role of knowledge evolution on improved decision-making.

Success Through Failure: The Paradox of Design by Henry Petroski - Using forecasting to create success by averting failure.

Adapt: Why Success Always Starts With Failure by Tim Harford - The importance of adaptive trial and error driven from the ground up rather than from the top down.

Charles
Received on Wed 27 Feb 2013 10:23:06 AM CST