VIC363H1 Chapter Notes - Chapter 2: Observer-Expectancy Effect, Selection Bias, Randomized Controlled Trial
Document Summary
Carole j. torgerson and david j. torger- son - Dominant paradigm in educational research = based on qualitative. Important questions about outcomes in education also need to be asked. Randomized controlled trial (rct) is rarely used in british educational research. Educational research community has focused on the development of non- experimental research methods. Few randomized trials of educational innovations have been undertaken. When interventions are not randomly allocated to schools: possible that the strategy actually had an educationally important effect but change was masked due to selection bias. Cannot make accurate comparison without a control group for reliable comparison. Non-randomized designs = matching schools with similar pupil intakes and past examination performance. Randomized designs give a more precise and robust estimate of effect. Potential confounding" variables: differential regression to the mean effects, parents of the intervention group children may have given more input to their children"s reading. Non-randomized quantitative methods are nearly always inferior to the randomized trial.