Between Subjects Methods

Question 1

Why experimental methods?

Answer 1

• allow us to demonstrated causality
• whereas observational research can only do correlations (directionality and third variable problem)

Question 2

How does experimental research work?

Answer 2

• Manipulate IV
• Measure DV
• Control extraneous variables to ensure that they don’t become confounds (if they vary systematically with the IV and affect the DV)
• Compare scores between conditions / treatments

Question 3

What are the disadvantages of experimental design

Answer 3

• need to know what to manipulate and what to measure
• need to know what to control tightly (based on literature, potential confounds)
• need precise hypothesis

Question 4

What is a between subjects experimental design:

Answer 4

• independent groups
• each group only undergoes one condition
• different treatments are applied to different but equivalent samples, and comparisons are made between the groups
• given independent scores

Question 5

If there are two conditions in a between subjects what test would you use?

Answer 5

Independent samples t-test (mann-whitney)

Question 6

If there were more than 2 conditions in a between subjects design what test would you use?

Answer 6

One way ANOVA (Kruskal Wallis if non parametric)

Question 7

What are the advantages and disadvantages of a Between Subjects design

Answer 7

+
- Not influenced by time-related factors (history effects, maturation)
- not influenced by order effects (practice / fatigue)

• requires larger number of participants
• vulnerable to some types of confounds (individual difference, environmental variables, other threats) because there are differences between the characteristics of the individuals in your study

Question 8

How do you avoid assignment / selection bias?

Answer 8

Individual differences can become a confounding variable
- on accident
- due to the experimenter (biases about what conditions they want)
- by participants themselves (self selection)

Aim for groups composed of equivalent individuals
- random assignment of ppts to group –> potentially confounding variables are randomly distributed
- restricted randomisation (where you create equal groups and match for a potential confound, then block randomise them)
- you could also restrict range (e.g. age)
- or hold a variable constant (all male)

Question 9

Individual differences and variance

Answer 9

statistical tests assess ratio of between group variance to within group variance.
We want high variance between groups, low variance within groups, therefore minimizing individual differences can be advantageous (higher statistical power).

But this compromises generalisability (external validity)

Question 10

What are environmental threats to internal validity?

Answer 10

Different participants tested at different times or places:
- time of day
- treatment / testing location
- weather
- noise

To reduce these threats you can run participants at same time of day, use same room, etc

Question 11

What is differential attrition and diffusion?

Answer 11

Differential attrition: drop out rates are systematic / biased
diffusion: you may think you’re manipulating the variable but you are not because students have been talking about it

Question 12

What is compensatory equalisation and compensatory rivalry

Answer 12

• when participants in the not preferred treatment receive some kind of increased effort / compensation (i.e. the teacher knows they are in the harsh / punitive condition so they try extra hard with teaching) (by experimenters)
• compensatory rivalry: when the group tries to outperform the intervention group to compensate for receiving less desirable treatment

Question 14

What is resentful demoralisation?

Answer 14

They perform even worse because they know they are in the bad condition. So they give up. (experimenters and particpiants)