Blogs About Reading
Shanahan on Literacy
Literacy expert Timothy Shanahan shares best practices for teaching reading and writing. Dr. Shanahan is an internationally recognized professor of urban education and reading researcher who has extensive experience with children in inner-city schools and children with special needs. All posts are reprinted with permission from Shanahan on Literacy.
Isn't Independent Reading a Research-Based Practice?
Teacher question: Dr. Shanahan, I know that you don’t support independent reading at school. However, in my graduate program we are learning that research evidence shows that kids who read the most become the best readers. I don’t get why you don’t support this research-based practice.
In grad school my statistics professor had us analyze some research data. It revealed a close connection between the number of school library books and kids’ reading achievement. Makes sense, right? The greater the availability of books, the better the students would read.
Unfortunately, what the data showed was that the more books available, the lower the kids’ reading ability.
There’s a rousing headline for you: Cut school library budgets so kids can learn to read!
Beware of correlations.
Your professor shows you the relationship between amount of student reading and how well students read, and you assume one of the variables must cause the other.
But correlation does not mean causation.
Go to the Buzz Feed website and you can see how increases in ice cream consumption lead to murder, something about which we should all be concerned. Obviously, that’s silly. I eat ice cream all the time and I’ve never killed anyone.
That odd correlation results because ice cream sales and murder are both related to a third variable, outdoor temperatures. As weather gets hotter, people eat more frozen custard and get more violent (the latter two variables have nothing to do with each other, in spite of the high correlation between them).
In my library book data, the weird connection resulted — not because book availability injures reading — but because school libraries at that time were funded on the basis of reading achievement. The poorer a school scored in reading, the more library funding it received.
Poor reading caused library books!
Correlations don’t tell us about causation or about the directions of relationship.
Your professors are absolutely correct that there are a lot of correlational studies showing that the best readers read the most. That’s a fact.
But there are several possible interpretations of this correlation.
- The more reading practice kids get the more their reading comprehension improves.
- The kids who can read best will choose to read more than the struggling readers do.
- Both amount of reading and reading achievement are attributable to some third variable.
- The relationship is reciprocal — some of the correlation is due to reading ability leading to reading practice, and some results from practice leading to reading ability.
Scientists have long been aware of specious correlation and have worked out ways for sorting out this kind of thing.
The most obvious fix is to test the patterns experimentally. One can, for instance, try to get kids to read more and measure the changes, if any, in their reading comprehension. Or, conversely, we can improve kids’ reading ability, and monitor what happens to the amount of independent reading.
Mostly investigations have usually explored the impact of practice on comprehension outcomes and have not been terribly successful. The results have ranged from no improvement to extremely modest gains (NICHD, 2000).
That doesn’t mean that reading practice can’t improve reading achievement, only that the types of practice evaluated so far haven’t done so. Most such studies have looked at “sustained silent reading" (SSR), the practice of setting aside class time for kids to read self-selected books.
A major flaw in these studies has been a lack of measurement of amount of reading. Schools may provide free reading time, but that doesn’t necessarily increase the amount of reading kids engage in. We may just be trading of one kind of reading for another, and in at least one study, the assigned free reading time apparently discourage kids from reading on their own (Summers & McClelland, 1982).
We not only don’t know if increasing kids’ reading practice leads to more learning, we don’t really know if our methods for increasing kids’ reading practice leads them to practice more.
Increased practice may improve achievement, but it is not clear that we know how to increase practice.
Teachers and publishers often tell me that they have improved on SSR (e.g., by adding reading conferences, quizzes). So far, no one has conducted a study showing, unambiguously, that we can increase kids’ amount of reading, and that those increases, consequently, lead to higher reading comprehension.
Experimental research in other realms suggest that not all practice is equal (Ericsson, 1993). "Deliberate practice" seems to be particularly profitable. That is practice that is purposeful and systematic, requiring focused attention and that is conducted with a specific goal of improving performance. Practicing under the auspices of a coach seems to matter, too.
These sound less like free reading and more like the reading that a teacher assigns.
Another way to figure this out is to conduct longitudinal studies in which amount of reading practice and reading achievement are each measured multiple times. Instead of correlating those two things with kids at a single time, we can track the influence of each across development. It is possible, for instance, to connect the amount of reading practice fourth-graders engage in with their gains in reading achievement between fourth- and fifth-grade.
Such studies, however, have failed to show a clear connection between earlier reading practice and reading comprehension gains (Aarnoutse & van Leeuwe, 1998). For instance, one of these studies concluded: “Reading achievement at age 10 significantly predicted independent reading at age 11. The alternative path, from independent reading at age 10 to reading achievement at age 11, was not significant.” (Harlaar, Deater- Deckard, Thompson, DeThorne, et al., 2011, p. 2123).
That study was able to attribute differences in both reading achievement and reading practice to genetic influences.
Another of these longitudinal correlational studies concluded that, “the results show that it is children’s reading skills that contribute to their subsequent out-of-school reading habits rather than vice versa: the more competent the children were in sentence comprehension, text reading, and word recognition at the end of first grade, the higher the amount of book and magazine reading.” (Leppanen, Aunola, & Nurmi, 2005, p. 395).
This study did find that reading practice was related to later improvements in word recognition but not enough to affect the kids’ reading comprehension.
Similarly, with older students, Cain and Oakhill (2011) reported that reading practice had positive impacts on vocabulary, but not comprehension, and that practice was more attributable to attainment than the opposite.
An interesting idea from these studies is that practice effects may be exerted through two separate mechanisms: one is the amount of words processed (a true practice effect) and the other is through the exertion of the students’ choice to read (a motivation effect — the aspect of practice thought to be genetically heritable).
If practice effects are divisible in that way, it would mean that it can’t be captured entirely by requiring additional reading at home or school.
My conclusions from all of this?
Increasing students amount of reading may have positive impacts on at least some aspects of reading (e.g., word recognition, fluency, vocabulary). And, over a long enough period of time, it is possible that those foundational improvements would result in improved reading comprehension — though neither experimental nor longitudinal correlational studies have yet found such a connection.
The practice effects that have been found are pretty small, so if they do eventually result in better comprehension, that would likely take a long time, and those effects would probably be even smaller.
The more certain affect, according to these longitudinal correlational studies, is that reading achievement influences desire to read. We still lack experimental evidence of that, however.
Increased practice could lead to some small achievement gains, but that doesn’t mean we know how best to get kids to read more. Swapping one form of school reading for another probably isn’t the answer, especially given that kids exercise no reading choice in that scenario (and given how hard it has been to generate learning from independent reading during the summer when no school reading must be sacrificed to allow it; see Jimmy Kim’s research, for instance).
Independent reading at school is not a research-based practice.
Use school time to raise reading achievement and find ways to encourage kids to choose to read on their own.
Doug Fisher has had great success in getting inner-city kids to read at home by making texts available, allowing students to choose what they want to read, influencing those choices through teacher book talks, and providing opportunities for kids to share socialy their home reading at school (e.g., book clubs).
That approach, though not yet proven to work by experimental study, intrigues me because it has the possibility of both increasing amount of student reading while encouraging students to choose to read on their own (according to the research, that dual approach should be a real plus).
And, it would be doing this while preserving the maximum amount of teaching; an approach more consistent with research findings that show achievement to have a bigger impact on practice, than the opposite.
Doug estimates that his students get 15 extra days of teaching each year this way (e.g., since 30 minutes of free reading per day across a 180-day school year displaces that much instruction or deliberate practice).
Now please let me enjoy my ice cream in peace.
Do you think we just don't know how to really design research to get the answers we need? Or perhaps the answers are really not in research? Always question...