Manipulating data analysis to achieve statistically significant results that are actually meaningless
P-hacking (also called data dredging or significance chasing) is the practice of manipulating data analysis — by testing multiple hypotheses, selectively excluding data points, adjusting variables, or trying different statistical tests — until a statistically significant result (typically p < 0.05) is obtained. The resulting "significant" finding may be a statistical artifact with no real-world meaning.
The p-value threshold of 0.05 means that there is a 5% probability of obtaining the observed result by chance if the null hypothesis is true. When researchers run dozens of tests on the same dataset, the probability of finding at least one "significant" result by chance increases dramatically — testing 20 hypotheses at the 0.05 level yields an expected one false positive purely by chance. P-hacking exploits this mathematical reality.
P-hacking is not a marginal problem. A 2015 analysis estimated that the rate of false positives in published research may be far higher than the assumed 5%. Pharmaceutical research is particularly vulnerable because the financial incentives for positive results are enormous. When clinical trials are designed, analyzed, and reported by companies with billions of dollars riding on the outcome, the temptation to explore the data until a favorable result emerges is overwhelming. The practice helps explain why many drug studies that appear promising in clinical trials fail to replicate in real-world use.
