Random allocation: A method that uses chance to assign participants to comparison groups in a trial, e.g. by using a random numbers table or a computer-generated random sequence.6 Random allocation implies that each individual (or unit) entered into a trial has the same chance of receiving each of the possible interventions. Also called random assignment.

Random error: Variation in a sample that can be expected to occur by chance.4, 6 Confidence intervals and p values allow for the existence of random error, but not systematic errors (bias). Also called nonsystematic error, random variation.

Random sample: A sample of n participants (or objects) selected from a population so that each has an equal and independent chance of being selected for the sample.4 Distinct from randomization and random allocation.

Random-effects model: A statistical model used in meta-analysis that assumes the true effects are normally distributed.32 Both within-study sampling error (variance) and between-studies variation are included in the assessment of the uncertainty (confidence interval) of the results.6 When there is heterogeneity among the results of the included studies beyond chance, random-effects models will provide wider confidence intervals than fixed-effect models.6 See also fixed-effect model.

Randomization: The process of randomly assigning participants to one of the arms of a controlled trial.6 Ensures that participants have an equal and independent chance of being in each arm of the study. There are two components to randomization: generation of a random sequence and its implementation, ideally in such a way that those enrolling participants into the study are not aware of the sequence (concealment of allocation).

Randomized controlled trial (RCT): An experimental study (controlled trial) in which participants are randomly assigned to treatment groups (experimental and control groups).4, 11

Rate: The speed or frequency that an event occurs, usually expressed with respect to time.6 For example, a mortality rate may be the number of deaths per year, per 100,000 individuals.

Recall bias: Bias arising from errors in recollecting events due to failures of memory and looking at things “in hindsight,” with possibly changed views.6 This bias is a threat to the validity of retrospective studies.

Reference population: The population to which the results of a study can be generalized. See also external validity.6

Regression analysis: A statistical modeling technique used to estimate or predict the influence of one or more independent variables on a dependent variable, e.g. the effect of age, sex, and educational level on the prevalence of a disease.6 Logistic regression and meta-regression are types of regression analysis.

Regression toward the mean: The phenomenon in which the results observed are influenced by a tendency for groups to reflect the grand population mean value.52 Regression to the mean is problematic when one group is selected on the basis of extreme values, and the comparison group is not. This is a common issue with disease-state management programs, which select outliers in one time period but “regress” to the mean value in subsequent time periods.

Relative risk (RR): See risk ratio.

Relative risk reduction (RRR): The proportional reduction in risk in one treatment group compared to another.6 Calculation: RRR = 1 – risk ratio, usually expressed as a percentage. For example, if the risk ratio is 0.25, then the relative risk reduction is 1 - 0.25 = 0.75 or 75%.

Reliability: The extent to which an instrument, scale, or other type of measurement or procedure yields consistent and reproducible results.4, 6, 7 Reliability is context-specific rather than a property of an instrument under all conditions. Lack of reliability can arise from divergences between observers or measurement instruments, measurement error, or instability in the attribute being measured. See also consistency.

Reporting bias: A bias caused by only a subset of all the relevant data being available.6 Studies in which an intervention is not found to be effective are sometimes not published. Because of this, systematic reviews that fail to include unpublished studies may overestimate the true effect of an intervention. In addition, a published report might present a biased set of results (e.g. only outcomes or sub-groups where a statistically significant difference was found). See also publication bias.

Representative population (or sample): A population or sample that is similar in important ways to the population to which the findings of a study are generalized.4

Research report: A report of accelerated practical research studies about the outcomes, comparative clinical effectiveness, safety, and appropriateness of health care items and services; one of the products from the Agency for Healthcare Research and Quality’s Effective Health Care Program.15 The research is conducted by centers known as Developing Evidence to Inform Decisions about Effectiveness (DEcIDE) Centers, which are health research organizations with access to health information databases and the capacity to conduct rapid turnaround research. Also called new research report.

Research review: A comprehensive report based on available evidence (evidence synthesis) that evaluates benefits and harms of alternative interventions and indicates where more research is needed.15 The Agency for Healthcare Research and Quality’s Effective Health Care Program produces two types of research reviews: comparative effectiveness (or effectiveness) reviews and technical briefs.

Retrospective study: A study that looks backward in time at outcomes of interest that have already occurred before the study was initiated.6 Case-control studies are usually retrospective, cohort studies sometimes are, and randomized controlled trials never are. See also prospective study.

Risk: The proportion of participants experiencing the event of interest over a specified period of time.6, 54 Often referred to as the event rate (experimental event rate and control event rate), however these terms confuse risk with rate. Calculation: Risk = number of events or newly affected persons / total persons observed, expressed as a proportion or a percentage. For example, if the event is observed in 25 out of 100 participants, the risk is 0.25 or 25%.

Risk difference: The difference in size of risk between two groups.6 For example, if one group has a 15% risk of contracting a particular disease, and the other has a 10% risk of getting the disease, the risk difference is 5%. Also called absolute risk difference, absolute risk reduction, or absolute risk increase, depending on the circumstances.

Risk factor: A term used to designate a characteristic that is more prevalent among participants who develop a given disease or outcome than among participants who do not.4

Risk ratio (RR): The ratio of risks in two groups. In intervention studies, it is the ratio of the risk in the experimental (exposed) group to the risk in the control (unexposed) group.6 A risk ratio of 1 indicates no difference in risk between the two groups. For undesirable outcomes, a risk ratio of < 1 indicates that the intervention was effective in reducing the risk of that outcome (e.g., the event is less likely to occur in the experimental than control group). Risk ratio is calculated in cohort or prospective studies.  Also called relative risk.

Robust: A term used to describe a statistical method if the outcome is not affected to a large extent by a violation of the assumptions of the method.4