Where we test,
measure, and publish.

Sample: 152 Indian adults, ages 18 to 55, recruited via social media and referral networks between January and June 2026.

Instrument: 14-item online questionnaire covering sleep duration, subjective sleep quality, morning recovery (primary outcome), screen use in the hour before bed, prior sleep interventions, and self-assessed health.

Primary outcome: Morning recovery was assessed via a single validated item: "How often do you wake feeling fully rested?" Response options: Always, Usually, Sometimes, Rarely, Never. Respondents selecting Rarely or Never were classified as waking unrefreshed.

Analysis: Descriptive statistics with cross-tabulation by self-reported sleep quality category. No inferential statistics claimed at this sample size.

Limitations: Self-reported data; volunteer sample may skew toward health-aware populations; not nationally representative by age, geography, or socioeconomic status. Causal inference is not supported. Volume 2 (August 2026, n=707) has been published. The 2027 edition targets n=1,000+ with a nationally representative panel.

Ethics: Respondent participation was voluntary and anonymous. No personally identifiable information was collected.

Model type: One-compartment first-order oral absorption and elimination model. Not a clinical measurement; an educational approximation based on published pharmacokinetics.

Parameters: Base half-life 5.0 hours (population average). Oral absorption rate constant 5.0 per hour. Clearance modifiers applied multiplicatively: smoking 0.5x, hormonal contraception 1.9x, pregnancy 2.5x, hepatic impairment 1.5x, CYP1A2 inhibitors (fluvoxamine, ciprofloxacin) 3.0x.

Sleep depth model: N3 sleep depth approximated as a decaying sinusoidal function of hours elapsed since sleep onset, normalised to 100% at the caffeine-free peak. Caffeine suppression applied via a Hill equation with sensitivity parameter k50 (low=160, typical=100, high=60 mg/L-equivalent). Maximum suppression capped at 45% of baseline depth.

Uncertainty band: Reflects approximately 90-fold person-to-person variation in circulating caffeine on an identical dose observed in controlled studies. The band spans a fast-clearing (half-life x0.55) and a slow-clearing (half-life x1.75) variant.

Output metric: "Shallower-equivalent minutes" is the area between the two curves converted to minutes of baseline-depth sleep. Nobody is awake for these minutes; the sleep is simply less deep than it would have been.

Caffeine values: Indian drink values sourced from published food composition data and manufacturer information. Filter coffee, tapri chai, masala chai, and major colas are included.

Primary citations: Baur DM et al. Caffeine and sleep. Journal of Sleep Research, 2024. Drake C et al. Caffeine effects taken 0, 3, or 6 hours before going to bed. Journal of Clinical Sleep Medicine, 2013.