Glossary

A   |  B   |  C   |  D   |  E   |  F   |  G   |  H   |  I   |  K   |  L
M   |  N   |  O   |  P   |  Q   |  R   |  S   |  T   |  V   |  W   |  Z

Glossary terms for 'N'

Nested case-control studyA case-control study in which the cases and controls are selected from a (larger) defined cohort or from among previously enrolled subjects in cohort study. This design is usually used when it is too expensive to make certain measurements in all of the subjects in the cohort; instead, they are made in samples that were stored at baseline in those subjects. For example, the investigators performed a nested case-control study to determine whether cytokine levels on newborn screening blood spots were associated with the development of cerebral palsy in the 2009 birth cohort of the state of Ohio.
Nominal variableA categorical variable for which there is no logical order. For example, religious affiliation (Christian, Buddhist, Hindu, Moslem, Jewish, other, none) was coded as a nominal variable.
Non-differential biasA type of bias that is not affected by whether a subject was a case or control (or occasionally, by whether a subject was exposed, or not exposed, to a third variable). For example, although recall of past exposure to antibiotics was imperfect in both cases and controls, the bias appeared to be non-differential, in that a review of medical records indicated that both groups had similar inaccuracies. Nondifferential bias tends to make associations harder to find because it reduces apparent differences between groups. See also differential bias.
Non-equivalence studyA study in which the investigator intends to show that two treatments do not have the same effect on the outcome. For example, the investigators did a non-equivalence study to show that 30-day survival among patients with pneumonia differed by whether they were treated with antibiotic A or antibiotic B.
Non-response biasA type of bias in which failure to respond (e.g., to a questionnaire) affects the results of a study. For example, the investigators were concerned about non-response bias in their study of the effects of illicit drug use on the risk of developing renal failure.
Null hypothesisThe form of the research hypothesis that specifies there is no difference in the groups being compared. For example, the null hypothesis stated that the risk of developing claudication would be the same in subjects with normal lipid levels who were treated with a statin as in those treated with placebo.
Number needed to treatThe absolute number of people who need to receive a treatment in order to prevent the occurrence of one outcome. Calculated as the reciprocal of the risk difference (see below). For example, when evaluating the benefits of treating mild-to-moderate hypertension, the number needed to treat was about 800 patients per year to prevent one stroke.

Glossary material from Hulley SB et al. Designing Clinical Research, 4th ed. Philadelphia, Lippincott Williams & Wilkins, 2013.