Prefer naturalistic over randomized controlled designs when studying dialogue-outcome relations¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article recommends examining naturally occurring variation in theoretically productive dialogue rather than imposing it via intervention, because interventions achieve only patchy uptake and control teachers may use the target forms anyway. It notes that "caution is already being advocated around the use of randomized controlled approaches in classroom research", and that in one cited study control classrooms averaged higher usage than the intervention group. The study therefore analyzed routine lessons and controlled identified confounds statistically.

Design Implications¶

Context¶

Requirements¶

• Naturally occurring variation in the dialogue features of interest across classrooms
• Identification and statistical control of confounding factors associated with both dialogue and outcome

Constraints¶

• Features not widely used must be discounted from naturalistic analysis
• The approach is particularly susceptible to confounding influences

Target Learners¶

• Primary school classrooms with teachers and students aged 10-11

Target Learning Goals¶

• Evaluating whether dialogue patterns relate to attainment, reasoning and attitudes

Related Strategies¶

• Use FOXGLOVE as a testbed for studying feedback uptake, variation, and downstream revision outcomes rather than rated quality alone
• Combine standardized tests and naturalistic classroom observations (mixed methods) when assessing oral communicative competence

Examples¶

-

Key Sources¶

• Howe, C., Hennessy, S., Mercer, N., Vrikki, M., & Wheatley, L. (2018). Teacher-Student Dialogue during Classroom Teaching: Does it Really Impact upon Student Outcomes? To appear in The Journal of the Learning Sciences. https://gtr.ukri.org/projects?ref=ES/M007103/1