Berkson's Paradox: Why Hospital Data Makes Healthy Smokers Look Fine
Berkson's paradox shows how selecting data from a biased pool can flip real-world correlations, and why your dataset's origin story matters as much as its contents.
C. Pearson
data science / statistics / probability
Mean Methods.
The mean is lying to you
n = 6 | p < 0.05 | σ ≠ 0
Latest
C. Pearson Archive
Lehmer's Trap: Why Random Number Generators Aren't Actually Random
Pseudorandom number generators follow deterministic rules, and that hidden structure can silently corrupt simulations, models, and statistical tests.
C. PearsonEndogeneity: The Reason Your Regression Coefficients Are Arguing With Themselves
Endogeneity corrupts regression estimates in ways that are hard to detect and easy to misinterpret. Here's what it is and why it matters.
C. PearsonOmitted Variable Bias: The Ghost Coefficient Haunting Your Regression
Omitted variable bias silently corrupts your regression coefficients when a missing variable correlates with both your predictor and outcome.
C. PearsonPoisson Processes: Why Rare Events Cluster When You Expect Them to Spread Out
Poisson processes explain why bus arrivals, server crashes, and earthquakes bunch together instead of spreading evenly across time.
C. PearsonAutocorrelation: Why Your Time Series Data Is Cheating on Your Model
Autocorrelation means your data points are secretly related to each other, and ignoring it makes your statistical conclusions quietly worthless.
C. Pearson EOF