|
Janet Looney reports back from the:
American Educational Research Association (AERA) Annual Meeting, held in San Francisco, USA, on April 7-10, 2006
The AERA annual meetings attract up to 15,000 attendees every year. The directory of presentations for the meeting is nearly 500 pages (size 8 font). It provides a “crash course” on developments in American educational research and policy. Needless to say, this report can only cover a sliver of the issues and topics discussed and debated at the conference.
A brief tour of some of the workshops I was able to attend:
- Racial diversity – Presenters pointed to rapidly growing minority populations in some parts of the U.S. (for example, a 400% increase in Hispanic population in North Carolina), and a real lack of research on issues related to Latino populations, on students with limited English proficiency, and so on. The Harvard School of Education civil rights project is worth visiting online (see www.civilrightsproject.harvard.edu/)
- Assessment of charter school performance - Presenters focused on the need to gather more specific and useful data to properly assess charter school performance. Contributors noted that we cannot assess the performance of charter schools (and thus their effectiveness in injecting competition and provoking wider-spread innovation in the education market) without also having some kind of baseline data from mainstream public schools.
- Large-scale interventions to improve student learning – Presenters discussed implementation and impact of different “whole-school designs” as an approach to large scale reform. Some whole school designs include detailed curriculum (there is no national curriculum in the U.S.), and some focus on school level reorganisation. Data on the impact of these designs on student achievement are mixed. The whole school designs are developed, copyrighted, and disseminated by for-profit and non-profit organisations.
- Effect Sizes and Confidence Intervals for Effect Sizes – this professional development workshop was exemplary on the “state of the art” in education research. The AERA has set up a panel to explore whether effect sizes need always to be reported (helping to compare “apples and oranges” when reviewing a range of quantitative studies). Apparently the American Psychological Association has held this standard since 1999. Currently, at least 24 American educational research journals require reports on effect sizes.
The No Child Left Behind Act (NCLB) is a major theme in American policy at this time. I attended a session which highlighted initial findings from a national longitudinal study of the NCLB and a study of state implementation of the NCLB. The research is being conducted by RAND and the American Institutes for Research (AIR). Of note for those interested in this subject:
- In 2003-04, 75 % of schools and 71% of school districts were making “adequate yearly progress” (AYP) as defined by the NCLB. The NCLB goals are for 100% of schools to reach this goal by 2013. The research shows that schools that are more diverse are less likely to make adequately yearly progress (39% of diverse schools vs. 10% of schools with a single sub-group are not reaching AYP goals). Most schools identified as not making adequate yearly progress are getting assistance. They are not being sanctioned.
- Schools not making AYP are required to provide extra tutoring for students. A market of for profit and non-profit providers is springing up to meet this need. Most are offering small group instruction (the presenters noted that this may be some cause for concern as they say small group instruction does not show positive results as compared to one-to-one instruction, which does, but which is rarely offered).
- NCLB allows parents to move their children to another school within the district if their school is not making AYP. School choice is often constrained by the absence of non-identified schools within a district. 20% of districts have no non-identified schools. 75% of districts in the nation have only one high school.
- The large majority of teachers across the country have been designated as highly qualified under NCLB, although many do not have a degree in the subject they are teaching. For example, 46% of highly qualified teachers in English do not have a degree in that subject. The percentage is even higher in mathematics.
A session on the What Works Clearinghouse (WWC), sponsored by the U.s. Department of Education, featured reviews on specific research domains which are priorities for the WWC (character education and early childhood intervention). At least for these reviews (and perhaps for all WWC research), programmes featured need to be branded and replicable (that is, have a name and be marketed to schools as a distinctly identifiable programme, practice or policy, and to be documented for replication purposes). The reviews privileged programmes with rigorous evaluation designs (i.e. Random Control Trials, and Quasi-Experimental Designs).
The presenters noted several challenges to identifying studies meeting these criteria, as well as in addressing:
- variations across programmes in the point of implementation (e.g. school-wide or in the classroom);
- questions regarding the fidelity of programme implementation;
- lack of data on programme impact for subgroups (e.g., different grade levels, students with special needs, or according to gender, ethnicity, language);
- whether randomization took place at the child, unit or school level (I learned elsewhere that currently no more than 5% of published articles in the U.S. use random samples).
|
