The variable manipulated by the experimenter. It is a feature of a task given to subjects or of the external or internal environment
The response measure of an experiment. It is the selected behavior which is measured to gauge the effect of the independent variable. The term also refers to the criterion variable, or Y variable, in a correlational study.
A procedure in which each member of the population has an equally likely change of being included in the sample.
Confounding (Nuisance) Variables
One or more independent variables that vary systematically with the variable of interest, decreasing the ability to make causal inferences.
A potential independent variable that is not to be manipulted in an experiment that must be neutralized to prevent confounding with the treatment variables.
Control (Placebo) Group
Group assigned to a reference, or baseline, condition consisting of the absence of a specific experimental treatment. Sometimes referred to as a placebo group when included in an experiment involving the administration of drugs.
Differences among the treatment means; reflects the effects of the treatments plus chance factors (experimental error)
A measure of variability based on the variaition of subjects treated alike; provides an estimate of experimental error
The differences among the treatment means in the population. A theoretical quantity that cannot be observed directly in an experiment.
Uncontrolled sources of variability (primarily individual differences) assumed to occur randomly during the course of an experiment.
The statistical hypothesis evaluated by hypothesis testing. Usually expressed as the absence of a relationship in the population and represented by the symbol H0.
The hypothesis that is accepted when the null hypothesis is rejected; represented by the symbol H1.
A statistical analysis involving the comparison of variances that reflect different sources of variability; abbreviated ANOVA.
The mean calculated from all the abservations in a study (ȲT)
Significance Level (α)
The probability (α) with which an experimenter is willing to reject the null hypothesis when in fact it is correct. Also known as the probability of a type I error.
Analytical comparisons specified before the start of an experiment
Comparisons not specified at the start of an experiment and conducted after the data have been examined. Also known as post hoc or multiple comparisons.
Type I Error
An error of statistical inference that occurs when the null hypothesis is true but is rejected. An error of "seeing too much in the data."
Type II Error
An error of statistical inference that occurs when the null hypothesis is false but is retained. An error of "not seeing enough in the data." Also notated by Beta (β).
Familywise Type I Error
The probability of committing a type I error over a set of statistical tests; approximately equal to the sum of the separate per comparison probabilities. Represented by the symbol FW.