Project 03: Lunar Effects on Biological Events

Source: bigastrologybook.com/2/research/19/project-3 Archive Date: 2026-03-21 Book: The Big Astrology Book of Research by Renay Oshop Dataset: 156,198,246 US births across 14,975 days (1969–2014)

Research Question

Do lunar phases — or the Moon's position in the zodiac — measurably influence the timing of spontaneous human biological events?

Hypothesis

The Moon's phase (especially the Full Moon at 180°) and zodiacal position trigger statistically significant deviations in the daily rate of biological events beyond what chance and known confounds would produce.

Why This Question

The "Full Moon Effect" is one of astrology's most culturally embedded claims — and one of the most persistently believed in medicine. ER nurses, obstetricians, and police dispatchers have long claimed that full moons bring chaos. Decades of studies have produced mixed results, but many suffered from the same flaw: failing to control for the powerful non-lunar rhythms already embedded in hospital and birth data.

This project addresses that directly with the largest dataset yet applied to the question.

Data

Field Detail
Dataset US daily birth counts, CDC (1969–1988, 1994–2003) + SSA (2004–2014)
Total events 156,198,246 births
Duration 45 years — 14,975 daily records
Astronomical data Swiss Ephemeris (pyswisseph) — lunar positions precise to <0.01°
Why births? Hospital admission data is often HIPAA-restricted or monthly-aggregated. Births provide a massive, high-resolution daily proxy for spontaneous biological activity, with public, verified data

Data is real and sourced from federal public health records. Not synthetic.

Methodology: The "Clean" Analysis

The central methodological challenge: birth data has powerful non-lunar rhythms that must be removed before any lunar signal becomes visible.

The detrending decomposition:

Observed = Trend × Seasonality × Weekly Cycle × Residual (Anomaly)

Each layer was stripped in sequence:

  1. Weekly cycle — weekends consistently show ~12–15% fewer births than weekdays (scheduled inductions and C-sections drive this). This is the strongest signal in the raw data and completely unrelated to lunar phase.

  2. Annual seasonality — US births peak in late summer (a well-known biological pattern). The "September Peak" and holiday depressions (Christmas: ~30% below average; Thanksgiving: ~35% below) are real and must be removed.

  3. Generational trend — the overall US birth rate changed substantially across 45 years. Long-term population dynamics were detrended separately.

What remains after stripping these three layers is the Anomaly Ratio: the actual birth count divided by what the model predicts based solely on week, season, and trend. An Anomaly Ratio of 1.00 means exactly as expected. This residual is what gets compared to lunar position.

Lunar variables tested:

Results

1. Lunar Phase: Null

Metric Result Interpretation
Full Moon (180°) ratio 1.001 +0.1% above average — noise
New Moon (0°) ratio 1.000 Exactly average
Maximum deviation observed +1.2% At random degrees, not at named phases
Visual pattern Flat No hump at 180°, no trough anywhere

The detrended anomaly line across the full 0–360° phase wheel is flat. There is no systematic elevation at the Full Moon, no systematic depression at the New Moon, no consistent pattern at any phase.

2. Sidereal Zodiac (Vedic/Lahiri): Null

Sign Position Ratio Interpretation
Aries (0°) 1.014 +1.4% — highest observed, but isolated and inconsistent
Libra (180°) 1.002 +0.2% — negligible
General trend Flat No sine wave, no consistent zodiacal signature

The Aries result (1.014) appears to be the kind of local fluctuation that will arise somewhere in any 360° scan of a noisy dataset. It is not reproduced consistently, does not align with a coherent zodiacal pattern, and does not survive scrutiny as a genuine signal.

3. Tropical Zodiac: Null

The Tropical zodiac results were statistically indistinguishable from the Sidereal results. Both lines trace random noise around the 1.0 baseline with no coherent structure.

4. Signal-to-Noise Quantification

Measure Value Meaning
Modulation Index 0.000001 Effectively zero coupling between moon position and birth rate
Rayleigh Test Technically significant But only because N=156M — any direction in a dataset this large registers as "significant"
Practical effect size Negligible <0.1% of birth variance explained by lunar position

The Rayleigh test result illustrates a critical statistical concept: with 156 million events, even an effect too small to mean anything in the real world will produce a statistically significant p-value. Statistical significance and practical significance are not the same thing. The modulation index of 0.000001 is the honest number — it tells you that lunar coupling, if it exists at all, is imperceptibly weak.

Why This Null Result Is Actually Strong Evidence

Many studies that claim to find lunar effects on hospital admissions fail to control for one or more of the confounds addressed here. When you don't remove the weekly cycle, you might find a "lunar effect" that is actually just a Monday effect. When you don't account for holidays, Christmas and Thanksgiving can look like they're driven by a different sky than normal days.

This study:

After all of that, the line is flat. This is not an absence of evidence — it is evidence of absence, at the scale that would matter for any claim about lunar influence on human biological timing.

Conclusion

After removing weekly rhythms, seasonal patterns, and generational trends from 45 years and 156 million births, the Moon's phase and zodiacal position show no measurable influence on when people are born.

The maximum observed deviation — +1.2% at a random phase angle — falls within the range expected from noise alone. The Full Moon ratio is 1.001. The Modulation Index is effectively zero.

The "Full Moon Effect" is one of medicine's most durable myths. This dataset is large enough to find it if it existed at meaningful scale. It didn't.

Archived code, raw data outputs, and visualization (clean_lunar_analysis_nobias.png) preserved in backup/.

Lunar Effects on Hospital Admissions

Research Question

Do lunar phases or zodiacal positions influence human biological events (births)? This project rigorously tests the "Full Moon Effect" and other astrological hypotheses using high-volume daily biological data.

Dataset

US Births (1969-2014)

Methodology: The "Clean" Analysis

To isolate potential lunar signals from strong human cycles, we employed a rigorous Detrending Decomposition:

$$ \text{Observed} = \text{Trend} \times \text{Seasonality} \times \text{Weekly Cycle} \times \text{Residual (Anomaly)} $$

  1. Weekly Detrending: Removes the ~15% drop in hospital activity on weekends.
  2. Annual Seasonality: Removes the "September Peak" and holiday impacts (e.g., Christmas drop) using a specific Day-of-Year baseline.
  3. Generational Trend: Removes long-term population growth and decline.

The Metric: The final analysis matches the Anomaly Ratio (Actual / Expected) against:

Execution

Technical Stack

Data Provenance

Hospital Admissions