Experiments: Extraneous Variables Flashcards
What are extraneous variables
Variables (not the IV or DV) that have not been controlled and that may have an impact on the dependent variable
What is a confounding variable
- a type of extraneous variable
- they’re variables that have varied systemically with the IV and therefore may be a reason for the change in the DV rather than the IV itself
- they happen in one condition and not the other = don’t know true impact of IV on DV
What are the 5 types of extraneous variables
- participant variables
- situational variables
- order effects
- Demand characteristics
- Researcher bias (Direct and Indirect)
What are participant variables
- characteristics of individual participants that might influence the outcome of a study
- e.g. gender, IQ, height, age
What are situational variables
- Any factor in the environment that could affect the DV
- e.g. temp, time, lighting, distractions
What are order effects + which design to they occur in
- Arises from order in which conditions (IV1 and IV2) are presented
- e.g. practice, boredom, fatigue
- they only happen in repeated measures design
What are Demand characteristics
- A cue that makes participants aware of the aims of the study or helps participants work out what the researcher expects to find
- this causes the participant to therefore change their behaviour
What is researcher bias
- Anything an investigator does that has an effect on the participant’s performance in a study
- this can be direct + indirect
What is direct researcher bias
- interacting with participants
- e.g. gestures/tone
What is indirect researcher bias
- how the study is designed
- e.g operationalising/duration
How can we control participant variables
- Through our samples (must be large + random)
- through our experimental design (repeated + matched pairs help)
- Through our allocation to conditions (randomly allocate participants to condition)
How can we control situational variables
Standardise
1. Standardised procedure (they do it when task is given)
2. Standardised instructions (written, recorded, printed…)
How do we control order effects
- use independent/matched pair design
- counter balancing (split sample in half. Half do conditions A then B, other half does B then A)
How do we control demand characteristics
Deception
- lie about aim
- filler Qs
- Distraction task
Experimental realism
- make task engaging so that participant pays attention to the task, rather than the fact they’re being observed
Single blind
- participant is unaware of what condition they’re in
How can we reduce researcher bias
Double blind
- Neither researcher or participant know which condition they’re in
Inter-rater reliability
- an independent researcher rates same behaviour at the same time as researcher,
- then they check for agreement (high positive correlation)
