Instructional Research and Research Synthesis Russell Gersten Instructional Research Group & University of Oregon
Focus Findings that challenge folk wisdom Multiplicities of interpretation The human factor (J. Singer)
Interventions for Struggling Readers: Torgesen Research Addressed common myth: work on word reading is all struggling readers need Most importantly, addresses a wild card: long term “fix” for struggling readers in grades 3-9 A professional issue for me since 1969
A few issues that bubble over Homogeneous groups seem to work best for building foundational skills/ Heterogeneous for comprehension The right texts for students to read as they work on comprehension Limits of the strategy instruction model (Beck & McKeown) & limited understanding of need to possess relevant background knowledge (before activating) We can only hope for increments
Carlo et al: Cognates Context: common feature of texts on teaching methods Most useful for technical words and Latin roots (for Latin languages) Ironies: English learners born in U. S. don’t have large tech vocabularies Catch 22: They don’t always work Does one teach awareness? Or really work on this ability
Thus the study Findings open up new issues Again, impact on words taught, NS on untaught cognates Implication: More to build broad awareness? Not a big deal? Deeper implications: Common finding for vocabulary instruction (depth, but not transfer)
The Research Synthesis: Part 1: Instructional Interventions Strength: Does not overshoot the data Points out that the adjustments are NOT CLEARLY EXPLAINED Helps point out that these need to be focus of next wave of instructional research: experimental and observational/ descriptive
The Research Synthesis: Part 2: Language of Instruction Politically controversial (unlike the first) Raises issue of how each meta-analysis is an act of interpretation and decision rules (defining terms, classifying interventions, methodological inclusion) affect outcomes
My Two Cents Worth RCT as gold standard is not enough Though some methodological issues stressed, others ignored Total confounding of teachers with intervention Differential attrition Use of covariates that are salient predictors (e.g. oral language has zero predictive validity) Control for instructional time
A very different picture would emerge