Docstoc

Experimental and Quasi-experimental Designs

Document Sample
Experimental and Quasi-experimental Designs Powered By Docstoc
					Experimental and Quasi-
experimental Designs

   Having fun with X’s and O’s
Inferring Causality

   Is the “effect” the result of the “cause”?
   Reality rarely accommodates us
   The mischief of intervening variables –
    threats to internal validity
       The year the Twins won the pennant
   Other threats to internal validity: p. 317, 328
Experimental Designs

   The language of experimental design
       X = intervention
       O = Observation
       O1 = Sequence of observation
       R = Randomization
       E = Experimental Group; C = Control
   The Gold Standard:
                              RE    O1    X   O2
                              RC    O1        O2
How a Classic Experiment Works (p. 324)

   Experimental Group                  Control Group
                          Compare:   Measure Dependent
   Measure Dependent
        Variable           Same?          Variable



    The Intervention




      Remeasure         Compare:        Remeasure
   Dependent Variable                Dependent Variable
                        Different?
Distinguishing Characteristic of
Experimental Designs
   Randomization: comparability of groups
Pre-Experimental Designs

   The Post-test only, or one-shot Design
       X    O1
   The One-Group Pretest-Posttest Design
       O1   X    O2
   Posttest only with nonequivalent groups (p.
    322)
                       X   O1
                           O2
Quasi-experimental Designs

   Classic Experimental Design, without
    Randomization


             E    O1   X     O2
             C    O1         O2
Beefing Up Quasi-Experiments

   Use Multiple Pretests
    O1   O2   O3    X       O4   O5   O6
Beefing Up Quasi-Experiments

   Vary the timing of the intervention



            E    O1    X     O2           O3
            C    O1          O2   X       O3
Beefing Up Quasi-Experiments

   Use non-equivalent interventions



             E    O1    X1   O2
             C    O1    X2   O2
Beefing Up Quasi-Experiments

   Select a Control group that matches the
    experimentals on variables of interest:


             EM   O1    X    O2
             CM   O1         O2

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:33
posted:8/1/2011
language:English
pages:11