Performance Validity (Effort) Testing

Description

Embed tasks so easy that anyone trying passes them, and treat scores below a cutoff, especially near chance, as evidence that a participant's whole session is invalid.

Clinical judgement is poor at detecting low effort; only specific validity tests work. These tests are free-standing or embedded in a battery and have a pass cutoff. In the Forced Choice Digit Memory Test a person sees a five-digit number and then picks it from a pair with a new number, over 32 trials; even severely amnesic patients score near perfectly unless distracted, so the cutoff is at least 90% correct, and scores near chance (50%) suggest deliberate failure. Green and colleagues found in 904 patients that low effort suppressed scores four and a half times more than moderate to severe brain injury, and that it affects the whole battery. Incentives drive failure. Person-fit statistics can additionally flag aberrant patterns such as inattention on easy items. Scores near the cutoff are ambiguous.

Attributes

Loss type
Unmigrated
Stage
Evaluation
Effort
Hours

Steps

  1. Choose or embed an effort measure that genuine participants pass easily, such as a forced-choice recognition task or trivially easy items within the main task.
  2. Set a cutoff (for example 90% correct) and note the chance level for the format.
  3. Administer it within the session, especially where incentives could motivate rushing, under-performance or over-performance.
  4. Score each participant; treat failures, and particularly near-chance scores, as invalidating that participant's other scores.
  5. Interpret near-cutoff scores cautiously and report exclusion rules and counts before analysis.
  6. Optionally compute person-fit statistics to flag aberrant response patterns.

Sources

  • Fundamentals of Human Neuropsychology (Kolb & Whishaw, 2021)

Also known as

  • Symptom validity testing
  • SVT
  • PVT
  • Effort testing
  • Forced Choice Digit Memory Test
  • Malingering detection

Categories

  • Research Operations
  • Research Design & Rigor