#3 Step 3 & 4: Measurement Reliability + Sampling Flashcards
What is reliability in measurement?
The extent to which a measurement is consistent
What is reliability across time + procedures to solve it?
We may not be able to get an identical score when we make measurements on the same person (because of the participant and environmental factors)
- Test-Retest Reliability: the extent to which scores on the same measured variable correlate with each other on the same measurement at 2 given times
- Parallel-Forms Reliability: the extent to which scores on similar, but not identical, measures administered at 2 different times correlate with one another
What is reliability across raters? How to improve inter-rater reliability?
We typically rely on 1 or more raters to quantify characteristics or behaviours.
- 1 rater: self-report
- more than 1 rater: observers, and could lead to random error (e.g. bias, mood etc)
Inter-rater reliability: extent to which ratings of one or more judge correlate
- train the raters
- revise scale to standardise
- have more raters to get aggregate
What is reliability across items and how to solve?
When the measurement has more than one item, there must be internal consistency (the extent to which items on a scale correlate with one another, thus measuring true score rather than a random error)
- split-half reliability
- Cronbach’s Coefficient Alpha
What are the 2 types of sampling?
- Probability sampling: the entire population is known, and the sampling process occurs by a random process based on probabilities
- Non-probability: the population is not completely known, and the method is based on common-sense or ease, with an effort to make representation
Types of probability sampling
- Random sampling
- Systematic sampling
- Stratified random sampling
- Proportionate stratified sampling
- Cluster sampling
Types of non-probability sampling
- Convenience sampling
- based on availability and willingness - Quota sampling
- identifying subgroup to be selected then establishing quotes (e.g. race)
Ethics: Belmont Report
- Respect for persons
- Beneficence
- Justice
Ethics: 4 Keys
- No harm
- Confidentiality
- Informed consent
- Debriefing