The research glossary defines terms used in conducting social science and policy research, for example those describing methods, measurements, statistical procedures, and other aspects of research; the child care glossary defines terms used to describe aspects of child care and early education practice and policy.
Attrition occurs when participants drop out of a longitudinal study or panel study over time. If particular types of study participants drop out at a higher rate than other types of participants, attrition can introduce bias that can potentially threaten the internal and external validity of a longitudinal study and the internal validity of an experiment with treatment and control groups.
A single value (mean, median, mode) representing the typical, normal, or middle value of a set of data.
Average Treatment Effect
The average treatment effect (ATE) measures the difference in the mean (average) outcome between the individuals or other units (e.g., classrooms, schools) assigned to the treatment and those assigned to the control. For example, in a study of the effects of a preschool reading intervention, the ATE would be the difference in average reading scores for children who received the intervention (treatment group) and the average reading scores for those who did not (control group).
A statement widely accepted as truth.
Bar charts are used by researchers to visually represent the frequencies or percentages with which different categories of a variable occur. They are most often used when describing and comparing the percentages of different groups with a specific characteristic. For example, the percentages of boys and girls who participate in team sports. However, they may also be used when describing averages such as the average time boys and girls spend per week participating in team sports. A bar is drawn for each of the groups along the horizontal axis and the height of the bar corresponds to the frequency or percentage with which the characteristic occurs (vertical axis).
Bayesian statistics is a general approach to estimating population parameters (characteristics) that uses both information about the prior distribution of the parameter of interest along with new evidence (likelihood function). In Bayesian statistics, the posterior probability distribution is the probability distribution once all information is taken into account.
A curve characteristic of a normal distribution, which is symmetrical about the mean and extends infinitely in both directions. The mean (average) is always in the center of the bell or normal curve. One half of the data points are to the left and one half are to the right of the mean.
The probability of making an error when comparing groups and stating that differences between the groups are the result of the chance variations when in reality the differences are the result of the experimental manipulation or intervention. Also referred to as the probability of making a Type II error.
A measure of the difference between the means of various groups.
Experimental design in which a different group of subjects are used for each level of the variable under study.