Control

1
Control

• Controlkey to internal validity
• Provides comparison and reduces variability
• Does not require a control group

2
Basic designs

• Between subjects/groups
• Within subjects/repeated measures

3
Control reduces variability
4
Strategies

• Setting
• Laboratoryhigh control
• Research preparationsparadigms (independent
  variable)
• Response instrumentation (Dependent)
• Assignment to conditionsavoiding subject
  variable confounds

5
Assignment to conditions
6
Within Subjects/Repeated Measures

• Each participant own control
• Equates groups
• Order/contrast effectspossible confound
• Decrease error variance
• Between Within

7
Within Subjects/Repeated Measures

• Limitations
• Lasting effects
• Carryover effects--counterbalancing
• Symmetric
• Asymmetric
• Eliminate individual differences

8
Between Subjects/Groups

• Random assignment
• Versus random sampling
• Ensure equivalence of groups prior to independent
  variable
• Random number table
• Matchingone to one
• Ensure equivalence on matched variable(s)
• Form pairs, triads, quartets, etc., on matching
  variable and randomly assign to group
• Matching variablerelated to dependent variable
• Multiple matching variables
• Equatingcharacteristics equally distributed
  across groups

9
Nuisance variables

• Cant be removed made an independent variable
• Eliminateuse only one level
• Reduces generalizability
• Eliminates regression effects
• Matched subjectsanalyze matching subject
  variable
• Eliminate error variance
• Randomized block design
• Independent variablesmatching subject (blocking)
  and manipulated variable(s)
• Blockingnominal, stratifiedordinal,
  levelinginterval/ratio

10
Statistical Controlcovariance analysis

• Equate groups statistically
• Without covariate (non-significant)
• With covariate (significant)

11

• Replication
• Direct/exact
• Systematic/conceptual
• Review literaturemeta-analysis
• Participants
• Special populations
• Numberstatistical power

12
(No Transcript)