Teach Effectively!

Robert Slavin and colleagues reported that reading programs that provide extensive professional development on instructional strategies which promote student participation, strengthen phonics competence, and explicitly teach comprehension strategies are the best bets for improving reading achievement. The clearest examples of the programs that led to the highest achievement were Direct Instruction and Success for All.

Writing in the December 2009 issue of the Review of Educational Research, Professor Slavin and colleagues reported the results of their examination of 142 studies. They wanted to determine whether curricula, technology, instructional processes, or combinations of curricula and processes produce greater reading achievement. The curriculum group included core reading programs, such as Reading Street and Open Court Reading. The technology group included programs that employ computers or similar methods such as computer-assisted instruction, multimedia (e.g., Reading Reels, or Writing to Read). The instructional process group included approaches that provide teachers effective strategies for teaching reading, such as Peer-Assisted Learning Strategies (PALS) and Cooperative Integrated Reading and Composition (CIRC). The combined curriculum-and-instructional-process group included programs that function as core curricula and also provide detailed professional development about using instructional strategies, such as Direct Instruction and Success for All. The researchers separated the studies into two groups: those with outcomes at the (a) beginning reading level vs. upper elementary level.
Continue reading ‘Go for DI and SFA’

In an article entitled “The Relation Between the Type and Amount of Instruction and Growth in Children’s Reading Competencies” that is slated to appear in American Educational Research Journal, Susan Sonnenschein and colleagues reported that kindergartners who enter school with relatively higher competence in literacy benefit more from instruction that emphasizes extracting meaning from what they read but their counterparts who enter kindergarten with lower literacy competence benefit more from instruction that emphasizes decoding. As children progress through the elementary grades, however, the effects of different instructional emphases lessen.

Although this difference obtained even after the research team took into account other factors (child ethnicity, parents’ education levels), these other factors affect have clear effects on children’s literacy development. African-American children, even those who were reading above average at the end of kindergarten, lost ground compared to their White, non-Hispanic peers in the third and fifth grades. Also, although there were not direct effects for the educational backgrounds of the teachers (number of reading courses teachers reported having taken), third-grade teachers’ background in reading influenced the amount of time they spent on reading.

Sonnenschein and colleagues examined data for over 6000 children, making these findings relatively solid. They argue that their results indicate that teachers may not adapt instruction to fit learners’ level of literacy competence.

However, these results will probably come as little or no surprise to many people. First, we already had a plenty of evidence that (a) high-performers are likely to continue to outperform low-performers over time (“thems that’s gots gets”) and (b) ethnicity matters in reading outcomes. Because we hope that educators adapt instruction to learners’ levels of development, the conclusion that teachers may not be doing so should be moot; however, such a judgment is based on educators’ hopes, and those can easily be dashed.

A latent growth model was used to investigate the longer term efficacy of phonics and integrated language arts instruction as well as amount of such instruction on children’s reading development, using the nationally representative Early Childhood Longitudinal Study data set (kindergarten through fifth grade). Type and amount of instruction were derived from teachers’ ratings. Children’s entry-level skills and ethnicity were predictors of children’s reading scores at the end of kindergarten. Ethnicity and parents’ education level predicted rate of growth. Type and amount of reading instruction predicted children’s reading scores. However, effects for type of instruction were time-sensitive, occurring only in kindergarten and first grade. Although children benefit from instruction in decoding and comprehension skills, instruction may not be optimally tailored to children most at risk.

Sonnenschein, S., Stapleton, L. M., & Benson, A. (2009). The relation between the type and amount of instruction and growth in children’s reading competencies. American Educational Research Journal. Advance online publication. doi: 10.3102/0002831209349215

Zig Engelmann, progenitor of Direct Instruction (DI), has posted a video of a talk he gave earlier this month. The presentation is an explication of the underlying principles of DI, “Theory of Direct Instruction.”

In the presentation (video below the jump), Mr. Engelmann shows some of his chops from his undergraduate degree in philosophy. He starts with philosophers’ fundamental arguments and shows how those correspond (or don’t) with learning and teaching concepts. For example, as he works through John Stuart Mills’ five methods of induction from A System of Logic, he makes clear how each would apply to teaching. I suspect that this particular sequence will show many people why DI instruction (the examples used in the scripts, not the teaching behavior) is structured the way it is.
Continue reading ‘Engelmann explains’

Over on Britannica Blog, Dan Willingham has a new post entitled “What Makes a Good Fourth-Grade Reader? Knowledge.” Professor Willingham asks, “What makes for effective reading instruction?” and then answers, “A new study indicates that an important contributor is integrating material from other subjects into reading instruction.”

He’s talking about a recently released study by Wai Ming Cheung and colleagues from the University of Hong Kong. They examined predictors of reading literacy among fourth graders and found that “the most powerful predictor [of high outcomes] was the use of materials from other subjects as reading resources.”

This finding is consistent with points made elsewhere as well as here on TE: It’s not sufficient to teach decoding and abstract strategies. Kids need to read stuff! That means they need real content, and certainly one of the best sources of that content would be what they’re learning in other courses. It’s relevant, probably pitched at their level, etc.

Reading literacy of fourth-grade students in Hong Kong showed a remarkable improvement from 2001 to 2006 as shown by international PIRLS studies. This study identified various aspects of the teacher factor contributing to the significant improvement among students. A total of 4,712 students and 144 teachers from 144 schools were randomly selected using probability proportional-to-size technique to receive the Reading Assessment Test and complete the Teacher’s Questionnaire, respectively. A number of items pertaining to teachers’ instructional strategies and activities, opportunities for students to read various types of materials, practices on assessment, and professional preparation and perception, were found to be significantly correlated with the outcome of students’ reading literacy. Stepwise regression procedure revealed four significant predictors for students’ overall reading achievement. The most powerful predictor was the use of materials from other subjects as reading resources. Suggestions to improve quality of teaching of reading and further studies are made.

Cheung, W. M, Tse, S. K., Lam, J. W. I., & Loh, E. K. Y. (2009). Progress in international reading literacy study 2006 (PIRLS): Pedagogical correlates of fourth-grade students in Hong Kong. Journal of Research in Reading, 32, 293-308.

Link to Professor Willingham’s blog entry. Link to the abstract for the study by Professor Cheung.

US Secretary of Education Arne Duncan spoke to the National Education Association today, calling on its members to work with him toward the goal of ensuring that “every child in America is learning from an effective teacher—no matter what it takes.”

That’s a noble goal. In my view, it requires that educators shed their allegiance to theory and adopt effective teaching practices. That’s what it takes.

To be sure, there is a lot of emphasis on high-stakes testing. Mr. Duncan even discussed it as one of the four features of his plan for reforming education. The effective teacher idea is another of the four. High-stakes tests (which need reform of their own, in my view—that’s another post) are only part of the game. Another is using evidence to guide instruction.

But even in the absence of results on high-stakes tests and from pristine research projects, educators can ascertain whether they’re using effective methods. That is, we can devise our own means of ascertaining whether teaching practices are effective at a more micro level. In a lot of ways, the process is relatively simple:

1. Identify goals and objectives in objectively measurable terms. How will we know if the students have learned X?
2. Identify the skills and knowledge that students will need to demonstrate mastery of those goals and objectives. What is required to show that a student can competently and independently do X?
3. Devise teaching algorithms for leading students to acquire the skills and knowledge identified in (ii).
   • Model: Demonstrate the skills for the students or tell them the knowledge;
   • Test: Have the students perform the skill or state the knowledge;
   • Coach: Reinforce and correct their performance;
   • Practice to mastery: Have them perform the skill or repeat the knowledge until they are facile with it and can do it under different, increasingly more challenging conditions.
4. Assess students’ competence according to the goals and objectives specified in (i).

Sure, I’ve simplified it here. Sure, the goals and objectives would need to be integrated in a fashion consistent with an epistemology of various subject areas. And, you’ll have to cut me a little slack about my use of words such as “tell” and “correct”; telling would need to include providing materials to read, for example. But the idea is just about as simple as I’ve sketched it here.

The area of early reading has been mapped according to this perspective on teaching. We can say how we would recognize a competent reader (i), what the component skills are (ii), and how to teach those skills. But the to-be-learned material doesn’t have to be elementary level reading. If the area is chemistry and the objectives were conducting an experiment using electrophoresis, I think we could perform a similar analysis.

Of course, the real trick will be to get people on board with such thinking. Mr. Duncan seems to want to do something like that, but he’s talking about things at a much grander scale.

Remarks of Arne Duncan to the National Education Association—Partners in Reform

Under the auspices of the Best Evidence Encyclopedia, Bob Slavin and colleagues Cynthia Lake, Susan Davis, and Nancy Madden released an analysis of the research literature on methods for teaching students who are struggling to learn to read, “Effective Programs for Struggling Readers: A Best-Evidence Synthesis.” In the synthesis they report the results of their examination of nearly 100 studies that used randomized or well-matched control groups, lasted for at least 12 weeks, and employed trustworthy measures of outcomes. The results of their review, which include both effect sizes and narrative descriptions of the studies, provide valuable insight into effective methods for remediating reading problems.

Key Findings

Overall, 96 experimental-control comparisons met the inclusion criteria, of which 38 used random assignment to treatments. Effect sizes (experimental-control differences as a proportion of a standard deviation) were averaged across studies, weighting by sample size.

One-to-One Tutoring by Teachers: ES=+0.38 in 19 studies
• Reading Recovery: ES=+0.23 in 8 studies
• Other programs: ES=+0.60 in 11 studies

One-to-One Tutoring by Paraprofessionals and Volunteers: ES=+0.24 in 18 studies
• Paraprofessionals: ES=+0.38 in 11 studies
• Volunteers: ES=+0.16 in 7 studies

Small Group Tutorials: ES=+0.38 in 11 studies

Classroom Instructional Process Approaches (low achievers): ES=+0.56 in 16 studies
• Cooperative Learning: ES=+0.58 in 8 studies

Classroom Instructional Process Programs with Tutoring (Success for All, low achievers): ES=+0.55 in 9 studies

Instructional Technology (low achievers): ES=+0.09 in 14 studies

Salvin, R. E., Lake, C., Davis, S., & Madden, N. A. (2009). Effective programs for stuggling readings: A best-evidence synthsis. Best Evidence Encyclopedia: http://www.bestevidence.org/reading/strug/strug_read.htm

I took considerable pleasure today in reading an article by Maria Glod of the Washington Post in which she reported about plans to develop national standards in reading and mathematics for students in US schools. In her article, entitled “46 States, D.C. Plan to Draft Common Education Standards,” Ms. Glod described efforts by the governors of most US states to describe a framework of knowledge and skills that would characterize a high-school diplomate who is ready for the world of work or higher education or, ideally, both.

Forty-six states and the District of Columbia today will announce an effort to craft a single vision for what children should learn each year from kindergarten through high school graduation, an unprecedented step toward a uniform definition of success in American schools.

The push for common reading and math standards marks a turning point in a movement to judge U.S. children using one yardstick that reflects expectations set for students in countries around the world at a time of global competition. Today, each state decides what to teach in third-grade reading, fifth-grade math and every other class. Critics think some set a bar so that students can pass tests but, ultimately, are ill-prepared.
Continue reading ‘Progress on US standards’

Jennifer Jennings and Andrew Beveridge reported that exempting students from tests, a controversial practice sometimes employed with students with disabilities, may have deleterious effects on the performance of younger students with disabilities. Here’s the abstract:

Analyzing data from a large urban district in Texas, this study examines how high-stakes test exemptions alter officially reported scores and asks whether test exemption has implications for the academic achievement of special education students. Test exemption inflated overall passing rates but especially affected the passing rates of African American and Hispanic students because these students were more likely to be exempted. Furthermore, our results suggest that tested special education students in Grades 3 through 8 performed better academically than they would have if they were not tested. However, taking the high-stakes test provided no academic benefit to special education students in Grades 9 through 11.

I rarely work on topics related to high-stakes testing, so I am not well-enough informed to comment on this paper; however, I thought it was interesting enough to merit mention here. What do readers make of this finding?

Jennings, J. L., & Beveridge, A. A. (2009). How does test exemption affect schools’ and students’ academic performance? Educational Evaluation and Policy Analysis, 31, 153-175.