Module 5-Slides Flashcards
Purpose of experimental design
~Provide evidence supporting a CAUSE-and-EFFECT relationship between INDEPENDENT and DEPENDENT variables
~Protect against threats to design validities
Pretest-Posttest Control Group Deisgn
~2 or more independent groups (treatments)
~1 IV
~1 or more DV
~Random assignment
~Known as parallel group study
~Scientific standard for investigating cause and effect relationship
Validity
P-P cont’d
~Design is strong in internal validity
–>randomization controls for potential bias
–>threat NOT controlled for include but not limited to attrition
Posttest-Only Control Group Design
~Similar to P-P control group design
~Except NO pretest is administered to either group
~Used when pretest is impractical or contraindicated
Validity
Posttest
~Internal validity is strong
~Assume groups are equivalent due to random assignment
Factorial Designs for Independent Groups
(2-way factorial design)
~Incorporates 2 or more IV (e/IV is a factor)
~Subjects randomly assigned to various combinations of the levels of the variables
~Larger samples are needed
~Described by # of levels within each factor
One-way Repeated Measures Design
~Also called within-subjects design
~Subjects used as their own control
Crossover Design
~Participants are RANDOMIZED to a Tx sequence
~Used to control for order effects
~Should only be used when pts’ condition is stable
~Considerations for “washout period” when the Tx has cumulative effects
2-way design w/ 2 repeated measures
~Used in studies w/ more than one IV
~Assuming participants were measured Pretest (baseline), Tx - immediate posttest, and 6-mo posttest
Between-groups, within-group Mixed design
~2 IV
–>1 (timing) repeated across all subjects (pretest and posttest)
–>the other randomized to independent groups (Tx)
Quasi-Experimental Designs
~Similar to experimental (or True Exp.) designs, EXCEPT lack either:
–>random assignment or
–>comparison group or both
~More likely than experimental designs to have threats to internal validity present
One-group Pretest-Posttest Design (Time series designs)
~All subjects receive the same Tx
~No comparison group, which limits internal and external validity
~IV is TIME rather than intervention
Repeated Measure Design (Time series designs)
~Multiple measures of the DV are taken
~All subjects receive the same Tx
~No comparison group, which limits internal and external validity
~IV is time
Nonequivalent Pretest-Posttest Control Group Design
~Similar to Pretest-posttest control group design except NO random assignment
~Nonequivalent pretest-posttest design
–>intact groups
–>Subject preferences
~Threats to validity include interaction of assignment/selection with Hx and maturation
Intact groups
comparing patients having migraine with healthy control subjects