Week 1 Flashcards
What are some ways to control for extraneous (secondary) variables? (5)
- elimination
- constancy
- making it into a IV
- randomisation
- statistical adjustment
What are some sources of invalidity in experiments?
- proactive history
- retroactive history
- repeated testing
- statistical regression (often happens if sample uses extreme scores)
- maturation
- loss of subjects
- interaction effects
- error in measurement of the DVs
- experimenter bias
- error from statistical inference
What is proactive history?
subject variables: try to minimise through elimination, constancy or random allocation
What is retroactive history?
something that happens to people during the experiment
What is statistical regression?
Because of test unreliability, extreme scores will move closer to the mean if you re-test them.
How can we account for statistical regression?
Control groups can help surmount this.
What can help with maturation?
control groups
Why is taking into consideration why people are dropping out in experiments important?
If it is from the IV (for example, horrific side effects from the medication, or their memory has amazingly improved), then there will be a big bias in the outcome of you experiment.
If drop out is NOT due to the IV, then we can ______
fix this statistically.
Why are dependent measures important to take into account when generalising results?
Because you have used a valid operationalisation? Is there more than one way to measure?
Will an independant samples between subjects t test and an ANOVA give you the same results?
yes
If an independent samples between subjects t-test and an ANOVA give you the same results, why would we ever use ANOVA?
Because t-tests can only be used to compare two means
Rather than simply reporting the results of the ANOVA (group X performed significantly greater than group Y), what must also report if there is a significant interaction?
The direction of the effect.
