# Experiments and Randomisation

• Experiments & Randomisation
• Designing an experiment
• Manipulation: When the experimenter systematically changes the levels of the
• The goal of the experiment is to investigate whether the manipulation of the IV causes an effect in the DV
• independent variable (IV) across the conditions (groups/ levels) of the experiment.
• Dependent variable (DV): The variable that is measured.
• IVs and DVs
• IV = What is manipulated; the CAUSE we are investigating
• DV = What is measured; where the EFFECT we are investigating would occur
• True versus Quasi experiments
• True experiments
• Researchers can actively manipulate the IV and so assign participants to which condition they participate in (between-subjects) and/or assign the order that participants see the conditions (within-subjects)
• Quasi experiments
• researchers are unable to actively manipulate the IV, either because they are investigating naturally existing variables or because it would be unethical to do so
• Between- or Within-subjects
• Between-subjects
• Independent groups
• Unrelated
• Unpaired
• Within-subjects
• Dependent groups
• Repeated measures
• Related
• Paired
• influences your choice of test and the conclusions you draw
• Confound/ nuisance variables
• Confounds = Other variables that may influence the DV – these need to be controlled in the design
• 1.Variables that are related to the DV and are known to influence it
• Confounds/ nuisance variables are threats to internal validity
• Internal validity
• we have greater confidence that the IV was responsible for any changes/differences observed in the DV
• Randomisation
• Participants in between-subjects experiments will have a random and equal chance of participating in any condition
• By hand: e.g., drawing allocation out of a hat
• By computer: computer programme can assign participants
• Use a random number generator (e.g., random.org) to produce an allocation sequence
• Counterbalancing
• Alternating the order that participants experience the condition
• Experimental control
• Internal validity
• we have greater confidence that the IV was responsible for any changes/differences observed in the DV
• ?control helps our experiment to have high internal validity