Situational Authenticity

Situational authenticity is one of the two complementary qualities of authenticity in Bachman and Palmer's framework (Language Testing in Practice, 1996). It refers to the degree of correspondence between the characteristics of a test task and the characteristics of tasks in the TLU domain. Where the test setting, input, and expected response resemble the real-world situation the test claims to generalise to, situational authenticity is high. Where the test introduces task features that have no counterpart in the TLU domain, it is low.

The five points of comparison

Situational authenticity is judged across the same five categories Bachman and Palmer use to describe any task: characteristics of the setting, the test rubric, the input, the expected response, and the relationship between input and response. Authenticity is the match between the test profile and the TLU profile across these dimensions. A reading task that asks candidates to scan a university course catalogue for specified facts under timed conditions is situationally authentic against a TLU domain of academic enrolment; the same task is situationally inauthentic against a TLU domain of leisure reading.

Situational versus interactional

Situational authenticity is necessary but not sufficient. A test task can match TLU surface features and still fail to engage the candidate's language ability in the way real language use does. Bachman and Palmer therefore pair situational with interactional authenticity, the degree to which the task engages the candidate's language ability, topical knowledge, and affective resources. The two together replace the older binary of "authentic versus inauthentic" with a more diagnostic two-axis frame.

A reading test built from real newspaper articles (high situational authenticity) but answered by candidates who skim only the question stems and scan for keywords (low interactional authenticity) is the classic case where one quality holds and the other does not.

Implications for test design

Situational authenticity is the easier of the two to engineer because it lives at the level of task specification. Audit each row of the test specification against the TLU description and ask: do these features actually appear together in the target domain? Where they do not, the spec is leaking inauthenticity into the test.

For AI-assisted item generation situational authenticity is the more tractable target. A generator can be conditioned on TLU-domain exemplars; the cost of staying situationally authentic is largely upstream prompt-engineering work. Interactional authenticity is harder because it depends on how the candidate engages the task, not just on the task's features.

Key References

• Bachman, L. F. & Palmer, A. S. (1996). Language Testing in Practice: Designing and Developing Useful Language Tests. Oxford University Press.
• Bachman, L. F. & Palmer, A. S. (2010). Language Assessment in Practice. Oxford University Press.

See Also

• Authenticity: the parent concept
• Interactional Authenticity: the complementary axis
• Target Language Use Domain: the reference against which situational authenticity is judged
• Test Specifications: where situational-authenticity decisions live operationally