From: Investigating and preventing scientific misconduct using Benford’s Law
Likely Appropriate Data | Examples |
Datasets spanning several orders of magnitude | World country populations across time |
Data derived from natural phenomena | Mathematical and physical constants |
Data with a positively skewed distribution, where the mean is greater than the median | Much ecological data such as river lengths |
Sets of numbers derived from combinations or functions of numbers | Regression coefficients of papers within a journal |
Likely Inappropriate Data | Examples |
Sets of assigned numbers or those driven more by human than natural processes | Sample or participant identification numbers, house prices |
Data that does not span several orders of magnitude (although we may apply the law to subsequent digits to the first) | Human heights, some stock market indexes |
Data with an expected specific non-positively skewed distribution | Binomial survival probability of polar bears across seasons |