Suggestion should be used daily, because only a few days within a given period are truly effective for suggestion.

Approximately 300 people for 90 consecutive days, with the first 30 days as an exploration phase; each day at the same local time, perform the same type of suggestion task using parallel items of equal difficulty, and randomly arrange suggestion and neutral blocks daily to estimate the net daily effect. Simultaneously record wrist activity, light exposure, sleep logs, and collect salivary cortisol at waking, 30 minutes after waking, and before the task; subsample measure dim-light melatonin onset and home sleep EEG. Pre-register the range of period length, the threshold for effective-day effect, and the bounds for the proportion of few days; compare continuous rhythmic models with sparse window models, using the last 60 days for out-of-sample prediction. Project duration approximately 12–18 months. To confirm circadian causal effects, a separate crossover experiment with approximately 60 people under controlled sleep/measurement phase is needed.

Daily Suggestion Responsiveness Across 90 Days: Circadian Phase, Cortisol, Sleep, and Preregistered Tests of Temporal Windows

If stable, predictable few effective days are obtained, this would correct statistical models where suggestion efficacy is mainly continuously varying without day-to-day switching windows.

Joseph P. Green，俄亥俄州立大学利马校区，催眠反应研究; Karin B. Jensen，卡罗琳斯卡学院，暗示与疼痛研究

'Only a few days are effective' is stronger than general state fluctuations, but lacks the period length and the bound of 'few'; only by supplementing these parameters in advance can a distinguishable prediction be made.

30 days can be a pilot study but is insufficient to reliably confirm a monthly-scale cycle; fixed clock time does not equal fixed physiological phase, a single cortisol sample cannot characterize secretory rhythm, and an effective day cannot be defined as p < .05 on that day.