Circular Statistics for Personality by Birth Angle #

Research Question #

Using proper circular statistics, do zodiac signs show non-uniform distributions for individuals with specific personality traits or professions?

Hypothesis #

If astrological claims are valid, individuals in professions associated with specific traits (e.g., performers = extraverts) should cluster at zodiacal positions traditionally linked to those traits (e.g., fire signs).

Background #

The Gauquelin studies claimed to find planetary "peaks" in certain chart sectors for professionals. This research applies modern circular statistics to test whether such patterns exist using real celebrity data with profession as a personality proxy.

Data Sources #

• Celebrity Data: 211 verified birth dates from Wikipedia/public records
  • 100 Performers (actors, musicians, comedians) — extraversion proxy
  • 50 Scientists (physicists, mathematicians, inventors) — introversion proxy
    
  • 61 Writers (novelists, poets, playwrights) — mixed/control
• Population Data: CDC births by month (2010-2020) for baseline rates

Mathematical Methods #

1. Rayleigh Test: Tests whether circular data is uniformly distributed or clustered
2. Chi-Square Test: Compares categorical distributions (sign/element by profession)
3. Fisher Exact Test: Tests specific claims (e.g., fire signs = extraversion)
4. Circular Mean & Concentration (R): Measures direction and strength of clustering

Key Findings #

Conclusion: No evidence that zodiac sign predicts personality or profession.

Limitations & Confounding Variables #

1. Seasonal Birth Confound #

Critical Issue: Births are NOT uniformly distributed across the year.

• CDC data shows Rayleigh p < 0.0001 for seasonal clustering
• More births in August-September (Leo/Virgo), fewer in January-February (Capricorn/Aquarius)
• This means Virgos are ~20% more common than Capricorns in the population
• Any "zodiac effect" must exceed this baseline non-uniformity

2. Selection Bias in Celebrity Data #

• Celebrities are not representative of the general population
• Success factors (socioeconomic background, geography, era) may correlate with birth timing
• Historical celebrities may have different seasonal birth patterns than modern populations
• Some birth dates may be unreliable (especially for older celebrities)

3. Profession as Personality Proxy #

• Profession is an imperfect proxy for personality
• Not all performers are extraverted; not all scientists are introverted
• Success in a field depends on many factors beyond personality
• Career choice is influenced by opportunity, not just disposition

4. Sample Size Limitations #

• 211 celebrities is modest for detecting small effects
• Power analysis: Can detect medium effects (d ≈ 0.4), may miss small effects
• Unequal group sizes (100 vs 50 vs 61) reduce statistical power
• Multiple comparison problem: 7+ tests increase false positive risk

5. Cultural & Historical Confounds #

• Birth timing patterns vary by era, country, and culture
• Holiday conceptions (December) → September births may vary historically
• Agricultural societies had different seasonal patterns than industrial ones
• Celebrity birth dates span multiple centuries

6. Zodiac Sign Assignment #

• Uses tropical zodiac (Western astrology standard)
• Sidereal zodiac would shift all signs by ~24°
• Cusp births (within ~2 days of sign boundary) may be misclassified
• No birth time data → cannot calculate rising sign or houses

7. File Drawer Problem #

• Null results are less likely to be published
• Our null findings are consistent with the published literature (Kelly 1979, Eysenck & Nias 1982)
• Significant results would require replication before accepting

Implementation #

Analysis Pipeline #

1. Parse celebrity birth dates and assign zodiac signs
2. Calculate CDC-based expected frequencies
3. Apply chi-square tests for sign/element × profession
4. Run Rayleigh tests for circular clustering by profession
5. Test specific claims (fire signs = extraversion)
6. Generate visualizations

Output Files #

• analysis_results.csv — Statistical test results
• celebrity_personality_data.csv — 211 celebrities with sign/element/profession
• circular_statistics_analysis.png — 4-panel visualization

Required Python Libraries #

pandas
numpy
scipy
matplotlib

References #

• Fisher, N.I. (1993). Statistical Analysis of Circular Data
• Kelly, I.W. (1979). Astrology and science: A critical examination. Psychological Reports
• Eysenck, H.J. & Nias, D.K.B. (1982). Astrology: Science or Superstition?
• CDC National Vital Statistics Reports (2010-2020)

Ethical Notes #

• Report null findings transparently
• Avoid deterministic claims about personality
• Acknowledge limitations openly
• Celebrity data is public information

Data Provenance #

Personality Simulation #

• Source: Synthetic/Generated Data.
• Description: Monte Carlo simulation of angular distribution.