Look up a term
The vocabulary, explained.
Definitions intended to assist in understanding an explanation. Use the related topic link or learn section to get detailed information on a complete calculation.
- Alternative hypothesis
- A population claim or set of parameter values compared with the null model.
- Bias
- A systematic tendency in measurement, sampling or estimation to miss the intended target.
- Binomial distribution
- A model for the success count across independent trials with a fixed trial count and a shared success probability.
- Confidence interval
- An interval produced by a procedure designed to cover a population parameter at the stated long-run rate.
- Confounding
- Distortion of an exposure–outcome comparison by factors associated with both the exposure and the outcome.
- Correlation
- A numerical measure of association. Pearson’s correlation coefficient measures the linear relationship between two numerical variables.
- Degrees of freedom
- The number of independent pieces of information available for a statistical calculation after accounting for constraints, such as fitted parameters.
- Effect size
- The size of a difference or association, expressed on an original or standardised scale.
- Expected value
- The average of a random variable’s possible values, weighted by their probabilities.
- Independent observations
- Observations whose joint distribution factors into their marginal distributions under the assumed model.
- Interquartile range
- For a dataset, the difference between the 75th percentile (third quartile) and the 25th percentile (first quartile). The interquartile range describes the spread of the middle half of the values.
- Likelihood
- The model’s probability or density for the observed data, treated as a function of its parameters.
- Mean
- The sum of the observations divided by how many there were or the expected value of a probability distribution.
- Median
- The middle point of a set of numbers arranged in order (from smallest to largest) that divides them into upper and lower halves.
- Null hypothesis
- The specific statistical model or assumption made for calculating the theoretical distribution of a test statistic.
- Odds ratio
- A ratio comparing the odds of an event in one group with its odds in another.
- P-value
- Probability of obtaining a test statistic as extreme or more extreme than the test statistic obtained when testing for compatibility with a null hypothesis (under the null).
- Population
- All possible units or all possible elements of the probability distribution from which inferences are being drawn.
- Power
- Probability of rejecting the null hypothesis under an alternative hypothesis and study design.
- Prevalence
- Proportion of a well-defined population having a particular disease status at a specific time or within a given time frame.
- Probability density
- A function used to compute the probability of continuous variables by determining the area under the curve of this function; it represents density at each point but does not represent a probability itself.
- Random sample
- Sample selected using some type of random process where sample statistics will be used to make inferences regarding some larger population.
- Relative risk
- The event risk in one group divided by the event risk in the comparison group, over a comparable follow-up period.
- Sample
- Observations actually sampled/available for statistical analysis.
- Sensitivity
- The proportion of people with the target condition who test positive.
- Specificity
- Among people without the target condition, the proportion testing negative.
- Standard deviation
- A measure of variation around the mean, expressed in the same units as the observations and calculated as the square root of variance.
- Standard error
- Standard deviation of a statistic's sampling distribution either derived empirically or assumed within some model formulation.
- Type I error
- Rejecting a null hypothesis that is actually true.
- Type II error
- Failing to reject a null hypothesis that is actually false.
- Variance
- The mean squared distance from a population mean, or a corresponding sample estimate using the stated denominator.
- Z-score
- Distance of a single value from its mean, in units of standard deviations.