Choosing the Right Designx

Report
Chapter 9
Chapter 9
Choosing the Right
Research Design
One-Way Designs
• The simplest possible experimental design
• Involves the manipulation of only one variable
(single independent variable)
One-Way Designs
• One-way designs must have a minimum of two
groups
• A two-groups design is the simplest type of one-way
design
• A one-way design with only two groups is most
often analyzed with…
• The Independent Samples t-test
One-Way Designs
• Experimental designs with more than two groups are
called multiple groups designs
• One-way multiple groups designs are most often
analyzed using…
• the one-way analysis of variance (ANOVA)
Factorial Designs
• When experimental designs involve more than one
independent variable they are called factorial designs
• Each independent variable has at least two levels (i.e.
conditions of the variable)
Factorial Designs
• Each independent variable is represented by
a separate number which indicates the
number of levels for that variable
• A 2 x 2 design has two independent variables
with 2 levels each
• A 2 x 3 x 4 design has three independent
variables. The first has 2 levels, the second
has 3 levels and the third has 4 levels.
Factorial Designs
• Factorial designs are most commonly
analyzed using…
• Univariate analysis of variance if only one
dependent variable is measured
• Multivariate analysis of variance
(MANOVA) for research with multiple
dependent variables
• 2 x 2 designs utilize a two-way ANOVA and
2 x 2 x 2 designs utilize a three-way ANOVA,
etc.
Factorial Designs
• There are 3 possible outcomes from a factorial
design:
• No significance
• Main effects
• Interactions
Factorial Designs
• Main effects indicate that a dependent variable is
significantly different across the levels of an
independent variable regardless of any other
independent variable.
• Interactions indicate that a dependent variable is
only significantly different across the levels of an
independent variable depending on the level of a
second independent variable.
Within-Subjects Designs
• Between-subjects designs include all of the
designs we have discussed so far
• Within-subjects or repeated measures designs
are those in which a participant serves in
more than one condition of a study.
Within-Subjects Designs
Advantages of within-subjects designs
• Fewer participants are needed because they are
used in multiple conditions
• Fewer participants are needed because the
design is more powerful
• There is less noise due to individual differences
• Thus person confounds are eliminated
• Within-subjects designs are the perfect form of
matching
Within-Subjects Designs
• Disadvantages of within-subjects designs
• Within-subjects designs are subject to certain
forms of bias:
• Sequence effects - when the passage of time between
conditions has an effect on performance
Within-Subjects Designs
• Disadvantages of within-Subjects Designs
• Carryover effects- when responses to one stimulus directly
influence the responses to another stimulus
• Figuring out the research hypothesis
Within-Subjects Designs
• Types of Carryover effects
• Order effects- when a question takes on a different
meaning following one question versus another or when a
stimulus is influenced following another stimulus
• Practice effects- when an experience with one task makes it
easier for someone to perform a different task
• Interference Effects- when an experience with one task
makes it more difficult for someone to perform a different
task
Within-Subjects Designs
• Solutions to problems of within-subjects
designs:
• Counterbalancing – Researcher varies the order in
which participants experience the experimental
conditions
• Complete counterbalancing – every possible order of
experimental treatments
• Reverse counterbalancing – create a single order and then
reverse it
• Partial counterbalancing - Selecting orders at random
Within-Subjects Designs
• Within-subjects or repeated measures designs
are most often analyzed using…
• Paired Samples T-test or
• Repeated measures analysis of variance
Mixed-Model Designs
• At least one independent variable is manipulated
between-subjects
• At least one independent variable is manipulated
within-subjects
• Mixed-model designs are analyzed using mixedmodel linear equation modeling

similar documents