Define the Control condition
Keep everything the same except the IV, it should be as similar to the experimental condition.
Define internal validity
Does ONLY the IV cause the change in the DV?
Define Noise
- Extraneous variable
- Unwanted variable causing variability within groups to increase e.g time of day/mood etc.
- Will not lead to a false effect.
Define confounding variables
- Third variable that has an effect on the relationship between the other variables
- Something that varies consistently/systematically with the IV
- Can be minimised through careful choice of comparisons
Ways to introduce confounding variables (HAM/HRR)
- History
- Attrition
- Maturation
- Hawthorne effect
- Researcher Bias
- Regression to the mean
Define Maturation
Ppts change in the experiment.
Define History
Something happens outside of experiment affecting the results. e.g safety and terrorist attack
Define the Hawthorne effect, and how to prevent this
- People expect their behaviours are measured so change their performance. Also known as Expectancy effect
- Prevent by using blind study: don’t know what condition they are in
How to stop Demand Characteristics from affecting exp
- Funnel Debriefing where you ask what the exp was about.
- Use implicit measures too as people can hide traits.
Define Researcher Bias
- When researcher acts in a way to cause bias. (Subtle cues)
- e.g if teachers expects child to do well, they will (Self fulfilling prophecy)
How to avoid expectancy effects and researcher bias
Double blind studies.
What are experimenter effects at analysis?
- Continuing data collection until p<0.05
Define HARKing
Hypothesising After Results are Known
Hypothesis should be clear before the data is collected
Define Regression to the Mean
People regress to the mean regardless e.g people who did bad on a test, when repeated that test would get higher regardless of teaching because they regress to the mean.
Define Attrition or Mortality
People who do not complete the study. Affects sample leading to a biased sample.
Define External Validity
- Do the observations generalise to populations/situations different to those in the experiment?
Types of generalisation (PEST)
- Population: is there something unusual about the population
- Stimulus: Is there something unusual about the experimental stimulus
- Environmental: Generalised to the real world? ecological validity.
- Temporal: Findings are valid as society has changed over time?
