Statistical Inference Vocabulary

a procedure for deciding between competing claims about a population

“Hypothesis testing revealed the drug was more effective than placebo.”

Origin: Greek hypothesis `foundation` + test

the default assumption that there is no effect or difference

“We reject the null hypothesis if p < 0.05.”

Origin: Latin nullus `none` + Greek hypothesis

the probability of observing results as extreme as the data, assuming the null hypothesis

“A p-value of 0.03 suggests the result is unlikely due to chance.”

Origin: p for probability + value

a range of values likely to contain the true population parameter

“The 95% confidence interval for the mean is 45 to 55.”

Origin: Latin confidere `to trust` + interval

the quality of being statistically unlikely to occur by chance

“The difference was statistically significant at the 0.01 level.”

Origin: Latin significare `to make known`, from signum `sign`

the probability of correctly rejecting a false null hypothesis

“We need a larger sample to achieve adequate statistical power.”

Origin: Latin potere `to be able`

incorrectly rejecting a true null hypothesis (false positive)

“Convicting an innocent person is analogous to a type I error.”

Origin: Statistical terminology from Jerzy Neyman and Egon Pearson

failing to reject a false null hypothesis (false negative)

“Missing a real disease is a type II error in medical testing.”

Origin: Statistical terminology from Jerzy Neyman and Egon Pearson

a subset of a population selected for analysis

“A random sample of 1000 voters was surveyed.”

Origin: Old French essample, from Latin exemplum `example`

systematic error that skews results in a particular direction

“Selection bias can invalidate survey results.”

Origin: French biais `slant, slope`, of unknown origin