Research Methods Experiments IV & DV Hypotheses (directional & non-directional) Controls Operationalisation psychlotron.org.uk
Experimental Design In an experiment we manipulate an IV There are usually two values of the IV e.g. Noise or no noise Rhymes or synonyms These determine the conditions of the experiment The conditions can be arranged in several different ways psychlotron.org.uk
Experimental Design Independent measures Repeated measures Matched participants psychlotron.org.uk
Independent Measures psychlotron.org.uk Recruit a group of participants Divide them into two This group does the experimental task with the IV set for condition 1 This group does the experimental task with the IV set for condition 2 Measure the DV for each group psychlotron.org.uk Compare the results for the two groups
Repeated Measures psychlotron.org.uk Recruit a group of participants Condition 1 Condition 2 The group does the experimental task with the IV set for condition 1 The group repeats the experimental task with the IV set for condition 2 psychlotron.org.uk Compare the results for the two conditions
Matched Participants psychlotron.org.uk Recruit a group of participants Find out what sorts of people you have in the group Recruit another group that matches them one for one Treat the experiment as independent measures Condition 1 Condition 2 psychlotron.org.uk Compare the results for the matched pairs
Participant Variables Variation between PPs can affect DV Could mask an effect (false negative) Could imply an effect where none exists (false positive) This is a problem with independent measures Control by random assignment to groups Use repeated measures or matched PPs instead psychlotron.org.uk
Fatigue, Boredom & Practice Carrying out a task repeatedly leads to changes in performance Deterioration as PPs become tired or bored Improvement due to practice This is a problem with repeated measures Leave a long gap between conditions Counterbalanced design Use independent measures or matched participants psychlotron.org.uk
Counterbalancing Important control when using repeated measures Reduces ‘carry over’ effects Half PPs do condition A then B Other half do condition B then A Fully counterbalanced: ABBA psychlotron.org.uk
Attrition When PPs drop out of a study (attrition) data are lost Fewer data = less powerful study This can be a problem with matched PPs (and repeated measures if there is a gap between conditions) Loss of a PP in these designs means losing their data from both conditions psychlotron.org.uk