Experiments with Two Groups

Experimental Psychology

Lecture, Chapter 10

Experimental Design

lExperimental design is the general plan for selecting participants, assigning participants to experimental conditions, controlling extraneous variables, and gathering data.

lThe Principle of Parsimony is the belief that explanations of phenomena and events should remain simple until the simple explanations are no longer valid.

Building an Experiment

lHow many IV’s? (1 IV)

lHow many groups? (2 groups, levels, or treatment conditions)

lExperimental group – in a 2 group design, the group of participants that receives the IV.

lControl group – in a 2 group design, the group of participants that does not receive the IV.

 Random Assignment

lRandom assignment is a method of assigning participants to groups so that each has an equal chance of being in any group.

lIndependent Groups are groups of participants formed by random assignment.

lDesigns with groups formed by random assignment are called independent groups designs, between-subjects designs, and/or randomized groups designs.

lWithout random assignment, experiments can be confounded, in that an extraneous variable systematically varies with the IV, thereby polluting validation of the cause-effect relationship.

Nonrandom Assignment

lNonrandom Assignment - a correlated assignment involves assigning participants to groups with an interrelationship, then randomly assigning to treatment conditions

lMatched pairs – pairing participants according to some relationship, then randomly assigning to treatment groups.

lRepeated measures – pairing participants with themselves, then administering both treatment levels

lNatural pairs – pairing participants according to a natural relationship, then randomly assigning to treatment groups.

Advantages of Correlated-Groups Design

lControl: methods for creating correlated-groups designs give us greater assurance of group equality and thus that the change in the DV is, in fact, due to the IV.

lStatistical issues: correlated-groups designs can help reduce error variation if matched on a relevant variable

lBetween groups variability (IV influencing DV)

lError variability (individual differences, measurement errors, and extraneous variation)

lDegrees of freedom (ability of a number in a specified set to assume any value

Advantages of Independent-Groups Design

lSimplicity

lStatistical advantage of correlated-groups design is mitigated as the sample size increases.

lCorrelated-groups designs are not always feasible

lConclusions:

Variations on the Two-Group Design

lComparing different amounts of the IV, rather than simply its presence or absence

lDealing with measured IV’s

lTrue experiments involve the direct manipulation of the IV (as well as random assignment)

lIn ex post facto research, the IV cannot be manipulated due to its nature; it is merely measured instead, limiting complete cause/effect conclusions.

Statistical Analysis of Two-Group Designs

lThe experimental design you use is related to the statistical test you use to analyze your data.

lAn experiment with 1 IV and 2 levels (groups) calls for a t test.

lIf participants are randomly assigned to groups, use a t test for independent samples.

lIf participants are correlated using matched pairs, repeated measures, or natural pairs, use a t test for correlated samples.

t Test for Independent Samples

In a computer analysis, you will receive the following information:

lDescriptive statistics, including # participants, means, SDs, and standard errors for both groups.

lInformation on homogeneity of variance

lt value and level of significance (should be < .05 if considered statistically significant).

t Test for Correlated Samples

In a computer analysis, you will receive the following information:

lDescriptive statistics for both groups, as in independent samples method

lInformation about the relationship between pairs of participants (should be highly positively correlated).

lt value and level of significance (should be < .05 if considered statistically significant).

Overview of Steps in Two-Group Designs

lAfter literature review, choose IV and DV.

lNarrow scope to one IV.

lChoose two levels of the IV.

lIf large # participants available, use random assignment, which yields independent groups. If small #, use correlated groups.

lIf independent groups, analyze with t test for independent samples; if correlated, analyze with t test for correlated samples.

lDraw conclusions