Glossary terms for 'T'
|T test (or Student?s t test)||A statistical test used to determine whether the mean value of a continuous variable in one group differs significantly from that in another group. For example, among study participants who were treated with two different antidepressants, a t test could be used to compare the mean depression scores after treatment in the two groups (an unpaired two-sample t test) or the mean change from baseline to after treatment in the two groups (a paired two-sample t test). See also one-sample tests. |
|Target population||A large set of people defined by clinical and demographic characteristics, to which the study investigator wishes to generalize the results of a study. For example, the target population for a study of a new treatment for asthma in children at the investigator?s hospital might be children with asthma throughout the world.|
|Time series design||A within-group study design in which measurements are made before and after each participant (or a whole community) receives an intervention. This design eliminates confounding because each participant serves as his own control. However, within group designs are susceptible to learning affects, regression to the mean, and secular trends. For example, fasting blood glucose levels might be measured among a group of patients with diabetes before starting an exercise program and again after the program has been completed to determine if exercise lowers fasting glucose levels. See also within-group design.|
|Translational research||Research that aims to translate scientific findings to improve health. Translational research may aim to test basic science findings from the laboratory in clinical studies in patients (often called “bench-to-bedside” or “T1 research”) or to apply the findings of clinical studies to improve health in populations (often called “bedside to population” or “T2 research”). For example, a study to determine whether a genetic defect that causes congenital deafness in mice has a similar effect in humans would be a T1 study; a study to determine whether a statewide effort to screen newborns with a test that measures cortical response to sound to detect hearing loss improves school performance would be a T2 study.|
|Type I error||An error in which a null hypothesis that is actually true in the population is rejected because of a statistically significant result in a study. For example, a Type I error occurs if a study of the effects of dietary carotene on the risk of developing colon cancer (with alpha set at 0.05) concludes that carotene reduces the risk of colon cancer (P < 0.05) when there is actually no association. See also false-positive rate. |
|Type II error||An error in which a null hypothesis that is actually false in the population is not rejected by a study (with P < alpha). For example, a Type II error occurs if a study fails to reject the null hypothesis that carotene has no effect on the risk of colon cancer (P > 0.05) when carotene actually does reduce the risk for colon cancer. See also false-negative rate.|
Glossary material from Hulley SB et al. Designing Clinical Research, 4th ed. Philadelphia, Lippincott Williams & Wilkins, 2013.