Experimental Design

• Experimental design refers to the way in which participants are arranged within different experimental conditions.
• Independent groups - participants are split into two groups and each only experience one level of IV.
• Independent Groups Evaluation

  • No order effects - Participants are less likely to get bored, tired or guess the aim as they have only experienced one condition.
  • Participants can confound the results - random allocation is used to deal with this.
  • Less economical - 2 times as many participants are needed to achieved this compared to repeated measures.
• Repeated measures - all participants experience all levels of IV and then the 2 mean scores from both conditions would be compared.
• Repeated Measures Evaluation

  • Participant variables are being controlled because the same participants are in both conditions.
  • More economical then independent groups or matched pairs as half the amount of participants are needed.
  • Order effects can impact the results - task completed in one condition, could have an impact on the task completed in another.
  • Demand characteristics can also be an issue
• Matched pairs - participants would be paired up based on a trait relevant to the study.
• Matched Pairs Evaluation

  • Order effects and demand characteristics are less of an issue cause participants only take part in one condition.
  • Impact of participant variables are reduced, but can never be eliminated completely.
  • Less economical than other designs, especially if a pre-test is required.
• Lab experiments are conducted in a highly controlled environment (not always a laboratory).
• Lab Experiments Evaluation

  • High control of EVs and CVs - allows use to determine cause and effect.
  • Increase replicability.
  • Lack of realism can increase demand characteristics.
  • Low external validity/Lacks generalisability.
• Field experiments - IV is still manipulated by the researcher, but in a natural, everyday setting.
• Field Experiment Evaluation

  • Higher mundane realism
  • Higher external validity
  • Less control of EV and CV - cause and effect are harder established.
  • Ethical Issues - consent and privacy - if they don't know they are in an experiment how can they consent.
• Quasi Experiment - IV is based on existing difference between. No one has manipulated this - it simply exists and cannot be changed. Eg; Age, Gender, Sex, Smokers ect.
• Quasi Experiment Evaluation

  • Often carried out in controlled conditions
  • Confounding variables can occur as use random allocation to conditions is not possible.
• Natural experiment - researcher has no control of IV and cannot change it. Would have occurred even if the experimenter was not studying it. IV is natural, not necessarily the setting. Eg; Natural disasters, system changes, technological development.
• Natural Experiment Evaluation

  • Opportunities for research that may not be conducted otherwise - for practical or ethical reasons.
  • High external validity.
  • Research worthy events many only happen very rarely.
  • It may not be possible to randomly allocate participants to different conditions.