Glossary of Common Research Terms

Terms found in the:

Introduction Section:

A Priori hypothesis: A hypothesis that is generated before the study or experiment. Not based upon experimental fact.

Methods Section:

Study Design

Before-After Trial: An investigation of therapeutic alternatives in which patients of one time period and under one treatment are compared with patients at a subsequent time who are treated in a different fashion. The same patients may be studied in the before and after periods, strengthening the power of the study to detect a difference between the groups.

Case-Control Study: Investigation generally used to test possible causes of a disease or disorder, in which persons with and without a designated disorder are compared with respect to prior or current exposure to a putative agent or causal factor. Often these types of studies are referred to as retrospective studies because the logic of the study design leads from effect to cause. For example, patients with hepatitis C (cases) and without hepatitis C (controls) are compared with respect to their exposure to blood products.

Case-Series: A series of patients with a defined disorder. Usually, the term is used to describe a study reporting on a consecutive collection of patients treated in a similar manner. For example, a surgeon might report on a series of 100 patients with carotid artery stenosis treated with a revascularization technique.

Cohort Study: A prospective investigation of the factors that might cause a disorder. A cohort (group) of persons who do not have the outcome of interest but who are exposed to the putative cause are compared to a cohort who are also free of the outcome but have not been exposed to the putative cause. Both cohorts are followed to compare the incidence of the outcome of interest. These studies are sometimes called prospective studies since the logic of the study design begins with exposure to the risk factor and the patients are followed for the development of the outcome. For example, a cohort of patients without evidence of liver disease but who have received a blood transfusion are compared to a cohort of patients without liver disease but have not been exposed to a blood transfusion. The two cohorts are followed over time and the incidence of hepatitis C is determined for each cohort.

Randomized Trial: An experiment in which subjects are randomly allocated (see random sampling below) to receive or not to receive an experimental, preventive, therapeutic, or diagnostic procedure and then are followed to determine the effect of the intervention.

Survey: Observational or descriptive, nonexperimental study in which persons are systematically examined for the presence or absence of characteristics of interest.


Consecutive Sample: Sample in which all the patients are chosen on a strict "first-come, first-chosen" basis. All persons who are eligible should be included as they are seen.

Convenience Sample: Persons or groups selected at the convenience of the investigator or primarily because they were available at a convenient time or place.

Random Sample: A sample derived by selecting patients such that each patient has an independent and fixed chance of selection. Whether or not a patient is selected is determined by chance, for example, by a table of random numbers.


Blind or Blinded: Clinicians or patients are blinded to the treatments the patients are receiving, or observers can be blinded to each other's assessments, making their observations unbiased by one another.

Double Blind or Blinded: Neither the subject nor the study staff (those responsible for patient treatment and data collection) is aware of the group or intervention to which the subject has been assigned.

Results Section

Confidence Interval: A measure of the statistical precision (stability of the estimate) of an observed effect size. It is usually expressed as the 95% confidence interval around the point estimate. For example, the effect of estrogen on the relative risk of endometrial cancer may be expressed as 7 (95% CI, 6 to 10), meaning that the relative risk interval between 6 and 10 has a 95% probability of containing the true estimate of risk

Likelihood Ratio: For a screening or a diagnostic test; expresses the relative odds that a given test would be expected in a patient with (as opposed to without) a disorder of interest.

Odds Ratio: A measure of the strength of association between having the risk factor if the disease or outcome is present compared to when it is absent. This type of measure is usually applied to cohort and cross-sectional studies.

P value: The P value is a quantitative estimate of the probability that the observed difference between two groups could have happened by chance alone.

Relative Risk Ratio: A measure of the strength of association between a risk factor and the outcome of interest, often a disease. It is calculated as the ratio of incidence of the outcome in outcomes in those exposed to the risk factor, compared to the incidence of disease in those not exposed to the risk factor. A ratio greater than one suggests increased risk, less than one a protective effect, and a ratio of one no additional risk.

Sensitivity: For a diagnostic or screening test, the proportion of people who truly have the designated disorder who are so identified by the test.

Specificity: For a diagnostic or screening test, the proportion of people who truly are free of a designated disorder, who are so identified by the test.