Description
Treats systematic disagreement between what users report and what implicit, physiological or behavioural measures show as a finding to explain, not as noise to discard.
Explicit self-reports and implicit measures often diverge: a participant reports feeling calm while electrodermal activity and hesitation spike, or rates a flow 'easy' after visibly struggling. The usual response is to pick the measure judged more valid. The mismatch approach instead treats divergence as information. Social desirability, politeness to the researcher, lack of interoceptive awareness, reconstructive memory and dual attitudes each predict different mismatch patterns. Analysing where, when and for whom measures diverge reveals hidden anxiety, face-saving or adaptation. Rigour requires time-aligned data across channels, pre-defined thresholds for what counts as divergence relative to each person's baseline, and follow-up probes such as cued retrospective interviews to interpret mismatches rather than speculating.
Attributes
- Discipline
- Neuro, Biometrics & Affect
- Loss type
- Latent
- Stage
- Evaluation
- Data
- Mixed
- Effort
- Weeks
Steps
- Choose paired measures for the constructs of interest, for example self-reported anxiety (a 1 to 7 rating) with electrodermal responses, or perceived ease (Single Ease Question) with task time and errors.
- Record baselines for each participant (for example 2 to 5 minutes of rest for physiology) and time-align all channels to task events.
- Standardise each measure within participant (z-scores against baseline) and define divergence, for example a self-report at or below the participant's median while the implicit measure exceeds +1 SD.
- Map mismatches by task moment and participant, and classify patterns: under-reporting, over-reporting, delayed response or segment-specific divergence.
- Run cued retrospective interviews, replaying the moments of mismatch to participants and asking what they were experiencing and why they rated as they did.
- Interpret patterns against candidate mechanisms (social desirability, low awareness, adaptation) and report mismatches as findings with design implications.
Sources
- Affective Computing (Picard)
- Methods Atlas — affective measurement (Gopinath)
- The Inquiry Series, Ch 37 — Picard's Affective Framework — Multi-Channel Emotion Measurement
Also known as
- Implicit-explicit dissociation analysis
- Say-feel mismatch
- Self-report versus physiology divergence
Categories
- Analytics & Behaviour
- Neuro, Biometrics & Affect