12 Feb 2012

Nature and Children’s Books Study Is For the Birds

Posted by the Writers of Sci-Why
The Huffington Post recently ran a story entitled, “Children's Books Lack Nature References, Study Suggests.” The study it referred to, which was published in the journal Sociological Inquiry, concluded that “today’s generation of children are [sic] not being socialized, at least through this source, toward an understanding and appreciation of the natural world and the place of humans within it.” (You can read the complete study here)
Here at Sci-Why, where we are both children’s writers/illustrators AND scientist/environmentalist types, we were naturally intrigued by this study. So we took a closer look at it.
As suspected, the study did not pass the scientific sniff test.
The study looked at 296 children’s books, published between 1938 and 2008, and which won the prestigious Caldecott Medal. The medal is awarded annually by the Association for Library Service to Children, a division of the American Library Association, “to the artist of the most distinguished American picture book for children.” From an examination of these books, the study authors drew conclusions about children’s books overall, and their effect on children.

In short, the conclusions the authors reach are not supportable by the facts, and the study’s design is flawed.

To begin with, the study examined books that are award-winners for artistic merit. These books, by definition, are not reflective of books overall.

Nor, as the authors claim, are Caldecott winners necessarily “the books that young children are most likely to encounter.”  Quoting a 15-year-old study, the authors say the Caldecott winners “are important both because the award leads to strong sales and they are featured in schools and libraries and influence tastes for children’s literature more generally.” While it may be true that Caldecott winners influence tastes in children’s literature, those tastes would be in artistic style, not in subject matter.
Furthermore, Caldecott winners are not necessarily the books children tend to encounter most. A better designed study would have looked at best-selling books, and books actually on school and library shelves. Caldecott winners are a tiny minority of these, and not reflective of them over all.
The choice of the sample, therefore, is seriously flawed. But an even greater flaw is the severely restricted size of the sample. The study examined just 296 books. Contrast this to the number of children’s books published in 2009, as reported by The Library and Book Trade Almanac (“Book Title Output and Average Prices:  2006-2009):” 21,878.
According to the American Library Association, that staggering figure is actually part of a downward trend in the number of published children’s books that began in 2008.
While we do not have access to the data describing the number of books published for children overall since 1938, considering the 2009 figure alone demonstrates the problem with this study. 296 books are simply too small a sample to reflect the nature of children’s books over all; in 2009, the Caldecott winner was just one title out of more than 20,000 books published for children in the U.S. One of of twenty-thousand yields a statistical correlation of exactly zero.
Furthermore, the authors' statistical analysis of trends over time is noted in three graphs. The data, the authors says, show statistical significance with a p-value of about .05. This is not a strong p-value. Something where p=.01 or less are stronger data.
Scientifically, then, the study fails to convince.  A non-scientific, common-sense approach to the study reveals even further flaws.
A quick eyeballing of the most popular picture books from the first “golden age of publishing” – a period loosely covering the 1950s and 1960s, features bestsellers such as Babar, Curious George, and books by Dr. Seuss and Richard Scarry. The Little Engine that Could. Madeline. Mike Mulligan and His Steam Shovel. Goodnight Moon. Where the Wild Things Are. 
Looking back even further, to the first illustrated children’s books in the 19th century, every single book features built environments, tamed nature and artificially civilized animals – think of Beatrix Potter’s  Peter in his little blue coat.
None of these books is “natural” in focus or illustration, yet they remain, perhaps, the most influential children’s books of all time.
In contrast, look at some notable books for 2012 from the Association for Library Service to Children (http://www.ala.org/alsc/awardsgrants/notalists/ncb). Nine out of 32 picture books on the list have natural themes and settings.
Interestingly, students of literature know that in the fairly short history of children’s literature, nature has rarely been presented as benevolent or even benign, making the current crop of books with pro-nature themes an anomaly. In traditional children’s stories, both oral and written, the wilderness is universally presented as a place of evil and danger. Consider Little Red Riding Hood or Hansel and Gretel as prime examples of the classic form.
There are still other issues with the study overall. For example, the authors use dated material, and also cite references that do not, in fact, support their claims. For example, consider this sentence: “the final decades of the twentieth century and the early years of the twenty-first saw a conservative backlash (Kline 2000; K. Gottfried, personal communication).” Unless the authors engaged in time travel, it would be impossible to draw conclusions about “the early years of the twenty-first century” from a document dated 2000.
Similarly, the authors of the study use data from 1996 to draw conclusions about the content of children’s science textbooks and continuing trends today. This data, 16 years out of date, does not reflect either the content of children’s texts today, nor the changing nature and usage of textbooks overall. To draw conclusions about books and their impact without consideration for the revolution in publishing we have been undergoing in the last decade is simply nonsensical.
The authors, we believe, suffer from a common problem among scientists and researchers of all stripes: confirmation bias. The researchers set out to confirm a hypothesis in which they already believed. Consider this quotation from one of the sources used in the study: ‘‘I believe one of the greatest causes of the ecological crisis is the state of personal alienation from nature in which many people live.’’  This is not a scientific observation; it is an unsubstantiated opinion.
As writers and illustrators involved in science and nature, we, of course, have our own biases. We know, however, exactly how much our own work, and the work of our peers, is inspired by and infused by the natural world. We see firsthand the work that is most influential, the work that is most often read by children and promoted by teachers and librarians.
Is it less “natural” than previous generations? We think not.
The Huffington Post should be more careful about the studies on which they choose to report. And for information about science or nature in children’s books, they should perhaps look to those who know something about the field: authors or librarians, not environmentalist-sociologists.


Sue Heavenrich said...

Too bad the authors of the study didn't take the time to visit their local library, eh? There's tons of good books with nature and environmental themes - so many that I can only review a fraction of them on my blog!

Janice Harayda said...

Thanks for your excellent analysis of why the study appears poorly designed. Its flaws may go beyond those you suggested. Recent children’s books have shown a trend toward what is often called “genre-bending,” or blurring the lines between types of books. That fluidity can make it harder to say whether or not a book is “about” something such as an appreciation of nature. And the influence of the Caldecott Medal may be waning in the digital age. Adults used to have far fewer sources of information about children’s books than they have had since the rise of online bookstores, networking sites, and book clubs. The profusion of those sources may dilute the impact of the medals.

But some of your points may confuse visitors to Sci-Why. You say: “The study looked at 296 children’s books, published between 1931 and 2008, which won the prestigious Caldecott Medal.” The Caldecott medals began in 1938 http://www.ala.org/alsc/awardsgrants/bookmedia/caldecottmedal/caldecottwinners/caldecottmedal, so a 1931 book could not have won. People may also feel confused by your statement that the “golden age of publishing” was one which loosely covers “the 1950s and 1960s,” and “features bestsellers such as Babar, Curious George, and books by Dr. Seuss and Richard Scarry. The Little Engine that Could. Madeline. Mike Mulligan and the Steam Shovel. Good Night Moon. Where the Wild Things Are.” Those books may have been bestsellers in the 1950s and 1960s, but “Babar,” “Madeline,” and “Mike Mulligan” first appeared in the 1930s and “Curious George” and “Goodnight Moon” in the 1940s. Finally, a couple of typos: The title of Margaret Wise Brown’s classic is “Goodnight Moon” (not Good Night) and that of Virginia Lee Burton’s is “Mike Mulligan and His Steam Shovel” (not the Steam).

None of these comments is intended to detract from your well-taken point that a “better designed study would have looked at best-selling books, and books actually on school and library shelves” or drawn on a wider sample. That is what Florida State University researchers did for study published in 2011 in Gender & Society that found that 57 percent of children’s books have male central characters and 31 percent have female central characters. They looked at almost 6, 000 books including Caldecott winners, the Little Golden Books series, and listings in Children’s Catalog http://oneminutebookreviews.wordpress.com/2011/05/07/the-glass-doghouse/.

Christine Tripp said...

A far more relevant study could and should have been conducted. A poll of randomly selected Teachers, Teacher Librarians, and Librarians, as to books being used in the Schools, requested and checked out by the Children, would have resulted in much different findings I am certain.
For this study to simply scan the Caldecott Winner titles listed on the Awards Site, has no merit. The Caldecott Honour Books for each year, are every bit as prestigious, and give a much different picture if added to the studies mix.
Just one example from a random year: 2010 Caldecott Honour book, "All the World" http://www.amazon.com/All-World-Liz-Garton-Scanlon/dp/1416985808

Ishta Mercurio said...

The title of the study outlines its primary flaw: they wanted to take what happens in award-winning books and then extrapolate. DUMB IDEA. As you point out, thinking that you can use less than 300 books to represent 70 years of publishing is lunacy. This study was screwed up even before they started it.

Why are so many scientists so stupid? And why do they get funding for this crap?