Calibration Certificate No. neglect-of-probability

Neglect of Probability

Neglect of probability is the tendency to respond to the emotional vividness or severity of a possible outcome while giving little or no weight to how likely that outcome actually is.

systematic

documented offset from the normative answer

Status — mixedgrounding ± moderate

The broad phenomenon — emotion-rich outcomes flatten sensitivity to probability — has converging support across the affective-risk literature, but the specific Rottenstreich & Hsee (2001) 'kisses' interaction replicates only weakly. A 2022 replication by Schulte-Mecklenbeck and colleagues (N=214, Swiss sample) found the affect-rich vs affect-poor interaction in the kiss-vs-money study but roughly halved (~20 points vs the original ~30+; chi-squared p≈0.05), and the interaction largely disappeared in the vacation-vs-tuition variant, with only main effects of outcome type surviving. Notably, Experiment 3 of the replication (shock vs cash-penalty WTP, the closest analog to the original $7/$10 finding) also showed no significant interaction between condition and punishment type (p=0.426), consistent with the failed Experiment 2 result. There is no large pre-registered meta-analysis dedicated to 'probability neglect' as a named effect; much of the policy-facing literature (Sunstein) is argumentative synthesis rather than new experiments.

Calibration record

Money, Kisses, and Electric Shocks~$7 vs ~$10 WTP across a 1% vs 99% probability gap for the shock (affect-rich); larger probability sensitivity for the affect-poor cash condition. Standardized effect sizes not reported.For affect-rich outcomes, willingness-to-pay was far less sensitive to probability: subjects would pay about $10 to avoid a 99% chance of a shock and about $7 to avoid a 1% chance of the same shock (only a ~$3 gap), whereas for the cash penalty WTP tracked probability much more closely. Weighting functions were more sharply S-shaped (over-weighting tiny probabilities, under-weighting large ones) for affect-rich than affect-poor outcomes. — Yuval Rottenstreich, Christopher K. Hsee, 2001 · Multiple student-sample experiments (per-study Ns on the order of dozens; exact Ns not reported in abstract/metadata).

Anticipatory stress under uncertaintyNull/near-null differences across probability levels; effect sizes not reported.Anticipatory arousal and reported tension showed essentially no differences across the event-probability conditions (except where probability was stated as zero) — physiological dread was driven by the prospect of the shock, not its likelihood. — Alan Monat, James R. Averill, Richard S. Lazarus, 1972 · Adult/student volunteers (N=120 male students; small per-condition Ns typical of early-1970s psychophysiology).

Source of systematic error

The dominant account is affective: when an outcome triggers strong emotion (dread, hope), attention fixes on the mental image of the outcome itself, and probability information is crowded out or processed only coarsely as 'might happen / won't happen.' Loewenstein, Weber, Hsee & Welch's (2001) 'risk as feelings' framework formalizes this — feelings experienced at the moment of decision are largely insensitive to probability variation, so when emotion runs high, behavior tracks outcome severity rather than likelihood. This is closely tied to the affect heuristic and to prospect theory's probability-weighting function becoming more extreme (steeper near 0 and 1, flatter in between) for affect-rich outcomes.

Alternatives frame the same data as probability-weighting distortion rather than wholesale 'neglect' (people do use probability, but with a strongly non-linear, emotion-modulated weighting function), or as a communication/format artifact (numeric probabilities are simply hard to evaluate and get rounded to qualitative categories). It is also debated whether 'probability neglect' is a distinct bias or a special case of the affect heuristic / risk-as-feelings.

Recalibration procedure

Force the probability to be quantified and put on the same scale as the stakes: state the risk as a frequency (e.g., 'X in 10,000 per year'), compare it to a familiar baseline risk, and compute an expected-value or 'cost per unit of risk reduced' before deciding. Sleeping on emotionally charged decisions and having a disinterested party restate the numbers also blunts the affect that drives the neglect.

Sunstein argues probability neglect is strongest under high emotion and recommends institutional constraints and explicit probability framing; Loewenstein et al.'s risk-as-feelings model implies that lowering emotional arousal or making likelihood concrete restores probability sensitivity.

Cross-calibrated against

Affect heuristicparentJudging risk and benefit by the emotional 'goodness/badness' tag attached to an outcome; probability neglect is often treated as what the affect heuristic produces when probability information competes with strong feeling.
Risk as feelingsmechanistically-linkedLoewenstein et al.'s (2001) framework supplying the core mechanism: decision-time feelings are largely insensitive to probability.
Denominator neglect (ratio bias)easily-confusedA different error: judging '1,286 in 10,000' as scarier than '24.14 in 100' (Yamagishi 1997). Here probability IS used but misread via salient numerators, rather than ignored.
Zero-risk biassiblingOver-valuing the elimination of a small risk to zero; relies on the same insensitivity to the size of a probability, focusing instead on the categorical shift to 'safe.'
Probability weighting (prospect theory)parentThe non-linear over/under-weighting of probabilities; probability neglect is the extreme, emotion-amplified end of this distortion.

Open the same case on your own draft.