External Validity
research-methodology
External validity is the degree to which research findings can be generalised beyond the specific participants, settings, and conditions of the study. A study with high external validity produces results that hold across different populations, contexts, and times.
Types
| Type | Question | SLA example |
|---|---|---|
| Population validity | Do findings apply to other learners? | Do results from university ESL students generalise to young EFL learners? |
| Ecological validity | Do findings apply in other settings? | Do lab results on Corrective Feedback hold in real classrooms? (See Ecological Validity) |
| Temporal validity | Do findings hold at other times? | Would a 1990s study on grammar instruction replicate in a 2020s communicative curriculum? |
| Treatment variation | Do findings hold with different implementations? | Does the effect of Focus on Form depend on the specific tasks used? |
The Tension with Internal Validity
Internal Validity and external validity often pull in opposite directions:
- Tight control (lab settings, homogeneous samples, standardised procedures) → high internal validity, low external validity
- Real-world conditions (diverse classrooms, varied teachers, natural constraints) → high ecological validity, weaker internal validity
No single study can maximise both. This is why Replication across contexts and meta-analysis across studies matter — they build external validity cumulatively.
Threats to External Validity
- Narrow sampling — most SLA research samples university students in Western contexts; generalisability to other populations is limited
- Volunteer bias — participants who volunteer may be more motivated
- Artificial settings — laboratory conditions do not represent classroom reality
- Treatment-setting interaction — a treatment that works in a resource-rich setting may fail elsewhere
- Pre-test sensitisation — pre-testing may prime participants, making results non-generalisable to unpretested populations (see Pre-test Post-test Design)
Enhancing External Validity
- Conducting research in authentic classroom settings (Classroom-based Research)
- Sampling diverse populations (different L1s, ages, proficiency levels, instructional contexts)
- Replication across multiple sites and conditions
- Meta-analyses synthesising findings across varied studies
- Transparent reporting of participant characteristics and setting details