CH 9

• Selecting and Recruiting Subjects


  The process of choosing participants for an experiment
• One Independent Variable: Two Group Designs

  • Two Independent Groups
  • Two Matched Groups
  • Multiple Groups
• Purpose of an experimental design

  Details an experimenter's plan for testing a hypothesis
• The design is the experiment's structure or floor plan—not the experiment's specific content
• We can use the same design to investigate different hypotheses
• What determines an experimental design

  • Number of independent variables in the hypothesis
  • Number of treatment conditions needed to test the hypothesis
  • Whether the same subjects are used in each treatment condition
• Between-subjects design


  A subject participates in only one condition of the experiment
• Representativeness of sample


  Determines whether we can generalize results to the entire population
• Random sampling


  Increases an experiment's external validity
• Minimum number of subjects per group

  10-20 subjects to detect a strong treatment effect
• Effect size


  Statistical estimate of the size or magnitude of a treatment effect
• Larger effect size


  Stronger relationship between IV and DV, fewer subjects needed to detect effect
• Two group design


  Involves creation of two separate groups of subjects
• Two independent groups design


  • One IV with two levels, subjects randomly assigned to conditions
• Random assignment


  Assigning subjects to conditions so each has equal chance of participating in each
• Purpose of random assignment

  To equally distribute subject variables between treatment groups and prevent confounding
• Experimental condition


  Presents a value of the independent variable
• Control condition


  Presents a zero level of the independent variable
• Experimental group-control group design


  Experimental group receives IV, control group receives no treatment
• Two experimental groups design


  Subjects assigned to one of two levels of the IV, randomization controls extraneous variables
• Random assignment works poorly with 5-10 subjects per condition
• Random assignment may not control all extraneous variables that could confound the experiment
• Two matched groups design


  Match participants on a subject variable correlated with DV, then randomly assign to conditions
• Purpose of matching

  To create equivalent groups on potentially confounding subject variables
• Precision matching


  Form pairs of identical scores on matching variable
• Range matching


  Form pairs of scores within a specified range on matching variable
• Rank-ordered matching


  Rank all scores on matching variable, form pairs of adjacent ranks
• When to use two matched groups design

  When there are two IV levels and an extraneous variable that could affect DV
• Multiple groups design


  • Between-subjects design with more than two IV levels
  • Multiple independent groups design: randomly assign subjects to treatment conditions
• Factors determining number of treatments

  Hypothesis, prior research, pilot study, practical limits
• Practical limitations on number of treatments

  Available subjects, time, expense
• Pilot study


  Trial run of experiment using a few subjects to refine procedure and determine if experiment is promising