PSYB01 - Chapter 10
Complex Experimental Designs
Increasing The Number Of Levels Of An Independent Variable
Experiment may with 3- more levels for several reasons .
Design with only two levels of independent variable cannot provide much info. about exact form
of relationship between independent and dependent variables .
Relationship is monotonic positive relationship, rather than strictly linear relationship .
Two levels only have a solid line .
Variables are sometimes related in curvilinear/ non-monotonic fashions .
The direction of relationship changes (e.g. Inverted-U) .
Experiment with only 2 levels of independent variable cannot detect curvilinear relationship .
At least need 3 levels in experimental design .
Increasing The Number Of Independent Variables: Factorial Design
Two/Three independent variables are operating simultaneously .
Closer to real-world conditions, which independent variable do not exist by themselves .
Factorial design: Designs with more than one independent variable .
All levels of each independent variable are combined with all levels of the other independent
Simplest factorial design: 2 X 2 factorial design, with 2 independent variables .
Format of factorial designs:
# of levels of first xV # of levels of second IVx Number of levels of third IV
Interpretation Of Factorial Designs
Factorial design yield 2 kinds of information:
1) Main effect : Information about effect of each independent variable taken by itself:
Two independent variables = 2 main effects .
2) Interaction: Interaction between 2 independent variables .
Effect of one independent variable depends on particular level of other variable .
Effect that independent variable has on dependent variable depends on level of other independent
Main effect: Effect each variable has by itself .
e.g. Type of question is overall effect of variable on dependent measure .
Main effect of each independent variable = Relationship between the independent & dependent