Description
Deliberately induce frustration, anxiety or monotony with rigged delays, goal changes or repetitive tasks, and measure physiological and self-rated responses at known moments.
Instead of waiting for emotions to occur, the researcher engineers them at logged moments. In the MIT frustration paradigm, participants play a timed, incentivised puzzle game under a cover story while the software randomly becomes unresponsive; synchronised skin conductance and blood-volume pressure are labelled by episode to train models distinguishing frustrated from non-frustrated states. The functional states variant induces anxiety (a ten-minute search task whose goal changes after three minutes, with time cut to five and then two minutes) or monotony (copying a whole website 'within an hour', stopped unannounced at 35 minutes). Attention-switching lability (Gorbov-Schulte table) and four 1-10 self-ratings (calm-anxious, energetic-fatigued, excited-depressed, confident-helpless) are taken before and after. Crossing these with product aesthetics and individual versus group testing shows how test conditions themselves distort usability results. Deception requires ethical approval and full debriefing.
Attributes
- Discipline
- Neuro, Biometrics & Affect
- Loss type
- Latent
- Stage
- Evaluation
- Data
- Quantitative
- Effort
- Weeks
Steps
- Choose the state to induce (frustration, anxiety, monotony) and the measures: physiological sensors, attention-switching tests or self-rating scales.
- Obtain ethical approval for deception, and prepare a plausible cover story.
- Take baseline measures before the task.
- Run the induction: random logged unresponsive episodes under time pressure and an incentive; or goal changes with time cuts (10 to 5 to 2 minutes); or a long repetitive task with an unannounced stop at about 35 minutes.
- Record synchronised physiological data or repeat the attention and self-rating measures after the task.
- Debrief every participant on the deception and offer to withdraw their data.
- Compare pre-post changes across conditions (for example design variants by individual or group testing), or train classifiers on episode-labelled data.
Sources
- Research Methods in HCI (Lazar et al., 2017)
- DUXU 2022: UX Research, Design, and Assessment (Soares et al., 2022)
Also known as
- Induced frustration with rigged system delays
- Frustration induction paradigm
- Frustrating the user on purpose
- Functional user state induction in usability testing
- Anxiety and monotony induction
- Functional states method
- Cumulative distortion testing
Categories
- Analytics & Behaviour
- Neuro, Biometrics & Affect