4 main problems with PURE OBSERVATION
- not always possible
- not always true
- can change over time
- not universal (opinionated)
How to deal with reliability of observations?
- Openness: Publicly available data
- Falsifiable hypotheses
- Double-Blind
Operationalization
Turn abstract concepts into measurable observations e.g. How to measure attraction?
Confirmatory vs exploratory studies
The former starts with a falsifiable hypothesis and data collected is used to confirm or contradict the hypothesis.
The latter starts off by collecting data and finding patterns to later form a hypothesis.
What makes a good instrument?
Validity
Reliability
Power/sensitivity
Why does correlation not mean causation?
Directionality and third variable problems
Demand characteristics
Characteristics that hint at or tell participants what a study is trying to observe or investigate.
Spurious correlations
Strongly correlated variables that we know are not causally related
Operational definition
Statement of procedures a researcher uses to measure a concept or variable.
What factors make a good operational definition?
Precise, specific and quantifiable.
