Ch. 5 Experimental Designs To Control Variables

Front 1

X

Back 1

Independent Variable that is manipulated

Front 2

O

Back 2

Process of observation or test

Front 3

R

Back 3

Random selection of subjects to groups

Front 4

Xs and Os across a given roq

Back 4

Apply to the Same people

Front 5

- - - - - -

Back 5

Dashed lines between groups indicate no random assignments to groups

Front 6

One-Shot case study design

Back 6

X O
A treatment is given to one group. Then observations (O) are made on the subjects in the group to test the effects of the treatment. The O are made in the form of the post-test.

No control group. Investigator makes inferences of the results based in what the results would have been if the intervention had not been given.

Weakest design. Think of the ex. given. What would the teach compare this too?

Note: Preexperimental design

Front 7

One-Group Pretest-Posttest design

Back 7

01 X 02
01=pretest 02=posttest
Group is administered a pretest is administered to measure the dependent variable and then the treatment is administered. The same test is readministered at the conclusion of the intervention (posttest).
Ex. The teach measured the students knowledge before the lesson and then tested there knowledge after.
Design doesn't control for student maturation, history, the testing situation, or statistical regression.
Considered to be a weak design, but is sometime necessary to use...like in schools that wont allow the students to be treated differently.

Note: Preexperimental design

Front 8

Static-Group Comparison Design

Back 8

X O
------------
O
Includes 2 treatment groups (experimental and control) in which each is only given the posttest. Subjects are not randomized.
Ex. Class 1 is control group and Class 2 is experimental group. Each class is given a posttest after intervention. Data is compared.
Design doesn't control for equality of the two groups. One class may have had more knowledge going into the experiment.

Note: Preexperimental design

Front 9

Posttest-Only Control Group Design

Back 9

R X 01
R 02
Includes 2 randomly assigned treatment groups. Randomly assigned Group 1 experiences intervention. Randomly assigned Group 2 does not. Both groups receive posttest.
Very powerful design because it controls for all threats to internal validity with exception to mortality.
No pretest used, useful because no interaction between pretest and IV, useful when pretest reactivity may occur.
Ex. Teacher would have to use a table of random #s to select students to select experimental and control groups. The experimental group would receive new curriculum. Control will be taught standard material. At the end each will be given posttest.

Front 10

Pretest-Post Test Control Group Design

Back 10

R 01 X 02
R 03 04
2 randomly assigned treatment groups. The first group is given a pretest then the intervention then a post test. The 2nd group is given a pretest then no intervention then post test.
Ex. Teach chooses 2 groups of students from large pool (individuals not classes). Each randomized group is given a pretest. The first group is then taught the experimental material. The second group is taught the non experimental material. Then both groups are given a post test.

Front 11

Solomon four-group design

Back 11

R 01 X 02
R 03 04
R X 05
R 06
Detailed designed that examines internal as well as external validity.
Examines multiple variables at once.

Front 12

T-test

Back 12

The t-test assesses whether the means of two groups are statistically different from each other. This analysis is appropriate whenever you want to compare the means of two groups, and especially appropriate as the analysis for the posttest-only two-group randomized experimental design.