# CI5390 Ch. 2 Randomized Experimental Designs Terms

## 27 terms

### analysis of covariance

A statistical method that compares two or more group means after adjustment for some control variable or covariate (such as pretest) to see if any differences between the adjusted means are statistically significant.

### class effects

The effects on students of being in a certain class.

### confounding

A situation in which the independent effects of two or more variables cannot be determined because the variables cannot be studied separately.

### continuous variable

A variable (such as age, test score, or height) that can take on a wide or infinite number of values.

### control group

A group assigned to be untreated or to receive a treament other than the experimental treatment.

### covariate

A control variable used in analysis of covariance or multiple regression analysis to adjust other variables.

### delayed treatment control group design

A design with a control group that will receive the experimental treatment later, after the study is over.

### dichotomous variable

A categorical variable (such as sex, on or off task, experimental control) that can take on only two values.

### disordinal interaction

An interaction between treatment and other variables in which the rank order of treatment groups depends on other variables.

### experimental comparison design

An experimental design that allows for the comparison of one treatment condition with another on two or more different groups.

### experimental group

A group assigned to receive some experimental treatment.

### experimental treatment

A treatment applied to some subjects in an experimental comparison design whose effects on one or more dependent (outcome) variable or variables are to be contrasted with the effects of other treatments or control (untreated) conditions.

### factor

A variable hypothesized to affect or cause another variable or variables; an independent variable.

### factorial design

An experimental comparison design in which treatments or other variables are analyzed as levels of one or more factors.

### Hawthorne effect

A tendency of subjects in an experimental group to exert outstanding efforts because they are conscious of being in an experiment, rather than because of the experimental treatments themselves.

### interaction

An effect on a dependent (outcome) variable of a combination of two or more factors or independent variables that is not simply the sum of the separate effects of the variables.

### main effect

A simple effect of a factor or independent variable on a dependent (outcome) variable.

### ordinal interaction

An interaction between treatment and other variables in which the rank order of the treatment groups does not depend on the other variables.

### posttest

A test or questionnaire given at the end of some treatment period.

### pretest

A test or questionnaire given before some treatment begins.

### random assignment

Selection into one or another treatment (or control) group in an experimental comparison design by chance, in such a way that all individuals to be assigned have a known and equal probability of being assigned to any given group.

### school effects

The effects on students or teachers of being in a particular school.

### selection bias

Any non-random factor that might influence the selection of individuals into one or another treatment.

### selection effects

Effects on outcomes of preexisting differences between subjects in experimental and control groups.

### statistical power

The ability of a statistical analysis to avoid getting a false negative error.

### stratified random assignment

Random assignment of subjects to one or more groups done in such a way as to ensure that each group will have certain characteristics.

### teacher effects

The effects on students of having a particular teacher.

Create Set