{
  "study_id": "NCT07133529",
  "agent": "baseline",
  "run": 1,
  "p_primary_hypothesis_supported": 0.25,
  "effect_direction": "favours intervention",
  "effect_size_90pct_interval": {"low": -0.2, "high": 0.5, "metric": "Cohen's d (within-subject, crossover) for uncertainty minus certainty in buffet intake (positive = more eaten under uncertainty)"},
  "moderator_prediction": "Stress eaters (high SSES) and those with larger cortisol/subjective stress responses to the uncertainty instruction eat more under uncertainty; low/under-eaters under stress may eat less, diluting the main effect.",
  "claim_ids_used": [],
  "rationale": "Pre-registered basic-science lab experiment (base rate ~44%). A mild informational manipulation (meal timing withheld) acting on an objective, highly variable outcome (ad libitum buffet intake), with all participants fasted and hungry in both sessions, so hunger likely swamps the certainty manipulation. Insurance-hypothesis and scarcity-cue work suggests a small positive effect, but such effects shrink in confirmatory tests. With n about 45 in a crossover, power for d_z about 0.2 is roughly 25-30%. Stress eating is bidirectional, so subgroup effects may cancel. Buffet intake is one of 28 primary outcomes, which lowers the chance it is emphasised or significant."
}