1. Chapter 11

2.  The general plan for carrying out a study where the independent variable is changed  Determines the internal validity  Should provide for limited differences between control and experimental groups

3.  Controlling threats to internal validity ◦ Preexperimental ◦ Quasi-experimental ◦ Experimental  Number of independent variables ◦ Single variable – one manipulated IV ◦ Factorial – two or more IVs & at lease one IV is manipulated

4.  One group pretest-posttest ◦ Give a pretest that measures the DV ◦ Administer the treatment ◦ Give a posttest that measures the DV ◦ Compare the pretest score to the posttest score ◦ Validity concerns  History  Maturation  Instrumentation  Regression

5.  Static group comparison ◦ Uses 2 or more preexisting intact or static groups ◦ One is exposed to the treatment ◦ The posttest is administered to both groups and the results compared ◦ Validity concerns  Non randomization – can’t generalize  Selection bias  Maturation  Mortality

6.  Nonrandomized control group (pre-post)  Counterbalanced design  Time-series with one group  Time-series with a control group

7.  Select two or more intact groups  Use similar groups  Randomly assign groups as experimental and control  Give pretest to both groups  Administer the treatment  Give posttest to both groups  Compare the test results and analyze  Validity concerns ◦ Selection bias can be controlled with the pretest and ANCOVA ◦ Interaction of selection and  maturation,  regression or  instrumentation

8.  Something in the selection process that causes one group to possess a higher or lower level of maturity ◦ Asking for volunteers

9.  May occur if each group represented a different population  ANCOVA with the pretest scores as the covariate is the best way to analyze

10.  Ceiling effect of tests could cause students with high scores on the pretest to show little improvement  students with lower scores on the pretest to show more improvement  ANCOVA with the pretest scores as the covariate is the best way to analyze

11.  Uses intact groups  Groups rotate positions  Contains a series of replications  Use when several treatments need to be studied  Rotation helps remove differences between groups. Experimental Treatments ReplicationX1X2X3 1Grp 1Grp 2Grp 3 2Grp 2Grp 3Grp 1 3Grp 3Grp 1Grp 2 Col. Mean

12.  Periodic measurement on one group  Introduces a treatment at some time period  Looking for changes in the pattern when treatment is added X

13.  Validity concerns ◦ History  Strengths-repeated measures help rule out these threats if they don’t cause fluxuations ◦ Maturation ◦ Testing ◦ regression

14.  Contains a control and experimental group  Helps control history effect  Can have multiple control and experimental groups

15.  Randomized Subjects, Posttest-Only Control Group  Randomized Matched Subjects, Posttest-Only Control Group  Randomized Subjects, Pretest-Posttest Control Group  Solomon Three-Group  Solomon Four-Group  Simple Factoral

16.  Randomly assign participants to groups(30 or more per group)  Give experimental group the treatment  Measure both on the dependent variable  All other variables are held constant  Uses randomization and a control group to control the threats to internal validity  Randomization ensures that initial differences between groups are due to chance  Controls for history, maturation, regression, pretesting

17.  Use for ◦ Studying changing attitudes ◦ Where pretests are not appropriate or available ◦ Can include more that 2 groups  Possible threats are subject effects and experimenter effects  Mortality could be a threat because there is no pretest to know if those that dropout are different from those that stay  Can’t be used to measure change

18.  Similar to previous design  Uses a matching procedure to create equivalent groups ◦ Matching variables should correlate to DV ◦ Matched pair must be randomly assigned one to the E group and one to the C. ◦ Could use a pretest in this model ◦ This procedure is useful for small groups (<30)  Threats to validity same as previous design

19.  Randomly assign participants to the experimental and control groups  Give a pretest on the DV to both groups  Administer the treatment to the experimental group  Give a posttest on the DV to both groups  Compare the pre and post test results using a t- test or F-test  ANCOVA is preferred statistic  Internal validity threat– test sensitizing  Main concern is with external validity involving interaction between the pretest and treatment.

20.  Use 3 groups  Random assignment  Same as randomized pretest-posttest, control group  Has second control group ◦ Not pretested ◦ Exposed to treatment  Compares all three posttest scores  Both control groups should be similar unless there is test sensitizing

21.  Use 4 groups  Random assignment  Same as randomized pretest-posttest, control group  Has third control group ◦ Not pretested ◦ Not Exposed to treatment  Compares all four posttest scores  Two groups take the pretest and two do not  Treatment is given to one pretest group and one none pretested group  Both control groups should be similar unless there is test sensitizing  Drawback is difficult to conduct and time consuming  Compare the posttests using ANOVA