Social Psychology Research

---

1/22/19

Correlational Research

• are x and y related?
• variables are just measured, not manipulated
• correlation ≠ causation
• spurious correlations = a third variable is causing the relationship
  • e.g., drowning and amount of ice cream eaten
• reverse causation/ bidirectional relations

Experiments

flowchart LR
	randomassignment["random assignment"]-->IndVar[treatment group]-->DepVar[outcome 1]
	randomassignment-->IndVar2[control group]-->DepVar2[outcome 2]

• Random assignment
  • Independent variables (manipulate)
    1. treatment group
    2. control group
  • Dependent variables (observe)
    1. treatment group → outcome 1
    2. control group → outcome 2
• keeping many factors constant → more precision

Evaluating Study Quality

• Objectivity – do you get the same result when someone else measures it?
• Reliability – do you get the same result when measured multiple times?
  • is a prerequisite for validity
    1. Internal validity – Did the independent variable really cause what you’re measuring?
    2. External validity – is the result generalizable?
• Validity – are you measuring what you want to measure?

Question

1. Explain what an experiment is and its crucial characteristics

Regression to the mean

• unusual values often produced by unusual circumstances or an unusual combination of factors that are unlikely to occur again

Statistical significance

• p-value: probability to obtain this result or a more extreme one if the null hypothesis is true
• significance is usually defined as
• depends on effect size and sample size