experimental designs

• Experimental design


  How the researcher uses the participant sample in combination with the different levels of independent variable
• Experimental designs


  • Repeated measures design
  • Independent groups design
  • Matched pairs design
• Independent groups design


  • Groups are independent of each other, different sets of people take part in each condition.
  • Participants are randomly allocated to each condition
  • Produces unrelated data.
• Evaluation of Independent groups design

  Strengths: avoids order effects (e.g. practice or tiredness) as people participate in one condition only.

  Limitations: more people are needed than with the repeated measures design.

  Individual differences between participants in the groups may affect results e.g. age or gender.
• Repeated measures design


  • The same participants complete in each of the conditions.
  • Data collected is related data.
• Evaluation of Repeated measures design

  strengths: participant variables are reduced as the same participants are used in each condition.

  fewer people are needed in each condition and therefore saves time.

  limitations: may be order effects since their performance could improve or worsen as they go on. however, can be controlled by counterbalancing.
• Counterbalancing


  Uses the ABBA format. Getting half the participants to do condition A first then B second and the other half B first then A second to balance the influence of order effects.
• Matched pairs design


  • Two separate groups of participants used, one for each condition
  • Participants ranked on a characteristic (e.g. aggression) and randomly assigned to conditions to balance that characteristic.
  • Data treated as related data
• Evaluation of Matched pairs design

  strengths: reduces participant variables as participants are matched based on relevant characteristics.

  no order effects as participants only take part in one condition.

  limitations: more time consuming since all characteristics have to be accurately matched.

  takes twice as many participants as repeated measures.

  participants are not identical so participant variables might still have effect.