Flashcards
5. Repeated Measurements
Cards in this deck (11)
- 1Purpose and implications of a within-subjects designEach participant is tested in each experimental condition (repeated measures).
- This means that fewer subjects are needed (as each subject receives all conditions).
- It also means that the equivalent groups problem is eliminated, increasing the power / sensitivity of statistical tests (since subjects serve as their own controls).
• Such tests basically compare the variability between conditions with the variability within conditions.
• Note: variability within conditions could be due to other systematic variance (extraneous factors), non-systematic individual differences or measurement errors. - 2When to use a within-subjects design- For some experiments within-subjects designs are the only reasonable choice, e.g. studies in which a trial lasts only a few seconds (e.g. reaction time), or when subjects are hard to come by (e.g. astronauts, patients, multilinguals).
• Suppose you are comparing RTs for two different traffic lights between subjects. If Group A seems to respond faster, individual differences may still have produced the result. If the same experiment is ran within subjects, individual differences can be eliminated from the estimate of variability between conditions. - 3Name two potential problems associated with a within-subjects design• Sequence or order effects are potential problems associated with this type of design.
• Another potential problem with within-subjects could be that being exposed to stimuli from all experimental conditions may lead to participants being able to ‘guess’ the purpose of the experiment - 4Explain why these are problems associated with a within-subjects design- Progressive effects: Performance may change due to practice, fatigue or boredom. In addition, spontaneous recovery may occur.
- Carryover effects: Systematic changes in performance that occur as a result of completing a sequence of conditions (easy hard vs. hard easy). - 5Two techniques by which to prevent order effectsCounterbalancing
Random order of trials - 6Explain the process of counterbalancing- Order effects can be controlled through counterbalancing:
With this method there must be two general categories, where participants are tested in each condition either just once or more than once.
• e.g. Reynolds (1992) studied the recognition of expertise in chess players, by testing once per condition. He utilized six games with about 20 moves.
With complete counterbalancing, every possible sequence will be used at least once. When many conditions are examined, complete counterbalancing is often impossible, e.g. with 6 conditions which would equal to 720 possible sequences.
• This confound could be broken by using 2 orders, where one-half participants carry out a ABCDEF sequence and one-half of the participants carry out a DBCFAE sequence. Thus, practice effects are balanced across conditions. - 7How to deal with participants 'guessing' the purpose of the experimente.g. a RT experiment testing recognition of words with positive (e.g. love) versus negative (e.g. death) emotional valence
One possible solution to this could be using ‘filler items’ as decoys (e.g. neutral words such as volume) - 8Strengths of a within-subjects designFew subjects needed
Equivalent samples
High statistical power - 9Strengths of a between-subjects designNaïve subjects
No order effects
Same procedure - 10Weaknesses of a within-subjects designAim may become obvious
Order effects
Procedure more complex - 11Weaknesses of a between-subjects designManu subjects needed
Equivalence problem
Lower statistical power
Information
- Description
- Psychology 2A (Research Planning and Design) Flashcards on 5. Repeated Measurements, created by Nina Olsson on 24/07/2018.
- Created by
- Nina Olsson
- Views
- 1
- Rating
More on Experimental and Quasi-Experimental Designs