affective-computing-basicslisted

## Overview affective-computing-basics covers the scientific measurement of emotion and affect: self-report instruments, facial expression analysis, vocal prosody, and peripheral physiology (EDA, heart rate, respiration). It focuses on validity — what each channel actually indexes — because emotion measurement is where convenient proxies most often replace the construct. The dimensional view (valence × arousal, plus dominance) organizes most measurement; discrete emotion categories ("basic emotions") are useful labels but contested as natural kinds. Measure dimensions; interpret categories cautiously. ## When to use - Choosing emotion measures for an experiment: self-report vs behavior vs physiology. - Facial expression analysis: FACS/action units vs black-box "emotion AI" classifiers. - Vocal affect: prosodic features (pitch, intensity, rate) and their limits. - Physiology: EDA (arousal), HRV (regulatory capacity), startle/facial EMG (valence). - Multimodal integration: combining channels without double-counting. - Evaluating commercial "emotion recognition" claims. ## Core concepts - **Self-report: still the gold standard for experience.** Validated scales (PANAS, SAM manikins, Geneva Emotion Wheel) measure felt affect directly. Limitations: introspection limits, demand characteristics, and retrospective bias — use momentary (EMA) over recalled reports when possible. Never treat a proxy as superior to self-report for subjective experience without evidence. - *