Experimental DesignChapter 8 Researcher manipulates the independent variable

Experimental Design-Chapter 8 Researcher manipulates the independent variable (usually to create groups) and then compares the groups in terms of their scores on the dependent variable (outcome measure) while keeping all other variables constant through direct experimental control or randomization- If score on the dependent variable are different then the researcher can conclude that the difference was due to the difference between groups and no other cause (and the experiment will have internal validity) pg 157 -8 A Confounding variable varies along with the independent variable. Confounding occurs when the effects of the independent variable and an uncontrolled variable are intertwined so you cannot determine which causes the effect

Basic Experiments The simplest experimental design has two variables, the independent and dependent with the independent variable having a minimum of two levels, an experimental and control group This type of experiment can take one of two possible forms 1) posttest only design or 2) pretest-posttest design Obtain two equivalent groups (random selection), introduce independent variable and then measure the effect of the independent variable on the dependent variable random assignment to groups or assign same subjects to both groups (CIT study with cross-over design)

Posttest only vs Pretest-Posttest design After groups formed (experimental and control) must choose two levels of the independent variable (treatment for the experimental group and no treatment for the control group) e. g. Experimental group gets treatment to stop smoking and control group does not Pretest-Posttest designs- the only difference between the posttest only and pretest-posttest design is that in the latter a pretest is given before the experimental manipulation is introduced

Posttest only vs Pretest-Posttest The pretest-posttest design makes it easier to assume the groups are equal at the beginning of the experiment. However if you have randomly assigned subjects to the different groups using a sufficiently large sample the groups should be equal without using a pretest Generally need a minimum of 20 -30 Subjects pg 160

Posttest only vs Pretest-Posttest advantages and disadvantages Advantages Pretest-Posttest While randomization is expected to produce equivalent groups this assumption may go unmet with small sample sizes and a pretest can increase the likelihood of equivalency Pretest may be necessary for assignment to groups so that those that score low or high on any pretest can be randomly assigned to conditions The comparison of pretest to posttest allows each subject to be evaluated in terms of change between the measures (with no pretest such comparison is not possible)

Posttest only vs Pretest-Posttest advantages and disadvantages Pretests helps determine the effects of attrition (dropout) –Can examine pretest scores of dropouts to determine if their scores differed from those completing the study Disadvantages Pretest A pretest may be time consuming A pretest may sensitize (alert) the subjects to the hypothesis which can result in changing a subject’s behavior in the study (can disguise the pretest as part of another study or embed the pretest in a series of irrelevant measures-time consuming)

Posttest only vs Pretest-Posttest advantages and disadvantages Solomon four group design- Half the subjects receive only the posttest and the other half receive both pretest and posttest. If there is no impact of the pretest, the posttest scores will be the same in the two control groups (with and without pretest) see table 8. 1 pg 162 Repeated measures has advantage of needing fewer subjects which decreases the effects of natural variation between individuals upon the results. Repeated subject designs are commonly used in longitudinal studies, over the long term, in educational tests where it is important to ensure that variability is low and in research on such functions as perception involving only a few subjects often receiving extensive training pg 164

Between group design vs. Repeated Measures design Between-group design is an experiment that has two or more groups of subjects each being tested by a different testing factor simultaneouslyeach subject is in either the treatment (experimental) group or the control group pg 163 A repeated-measures design is one in which multiple, or repeated, measurements are made on each subject. weekly blood pressures each subject measured after receiving each level of independent variable

Between group design vs. Repeated Measures design In the between groups design subjects are assigned to each of the conditions using random assignment https: //www. randomizer. org/ In repeated measures the same individual participates in all of the groups. These studies are more sensitive to finding statistically significant results-Even if you have randomly selected and assigned subjects to conditions in the between groups design there is still individual variation (naturally occurring “random error”-differences between the subjects assigned to the different groups) which may make the effect of the independent variable unclear but when testing the same person in different conditions (versus different persons in different conditions) this random error is eliminated

Between group design vs. Repeated Measures design One limitation of repeated measures is that the conditions must be presented in a particular sequence which could result in an order effect-the order of presenting the treatments affects the dependent (outcome) variable (maybe a subject performs better in the secondition because of practice in the first condition (practice effect) or performed poorer in the secondition due to fatigue (fatigue effect) or that the first treatment influences the second treatment (carryover effect) Carryover effect occurs when the first condition produces a change that is still influencing the person when the secondition is introduced

Repeated Measures-types of counterbalancing Counterbalancing All possible orders of presentation are included in the experiment. Latin Square-A Latin square is an table filled with n x n different symbols in such a way that each symbol occurs exactly once in each row and exactly once in each column. Each condition appears at each ordinal position (1 st 2 nd 3 rd etc. ) and occurs exactly once in each row and once in each column Using a Latin square controls for most order effects without having to include all possible orders (each condition precedes and follows each condition one time)

Matched Pairs Design Rather than using random assignment to groups you can first match subjects on a variable (achieving equivalency in this manner rather than through randomization) and avoid repeated measures/counterbalanced designs pg 169 Example study 1000 subjects each receive one of two treatments - a placebo or a cold vaccine. The 1000 subjects are grouped into 500 matched pairs. Each pair is matched on gender and age. For example, Pair 1 might be two women, both age 21. Pair 2 might be two men, both age 21. Pair 3 might be two women, both age 22 matched 1) matched

Activity Design a repeated measures experiment that investigates the effect of report presentation style on the grade received for the report. Use two levels of the independent variable; a “professional style” presentation (high-quality paper, consistent use of margins and fonts, carefully constructed tables and charts) and a “nonprofessional style” (average-quality paper, frequent changes in the margins and fonts, tables and charts lacking proper labels). Discuss the necessity for using counterbalancing.