Lecture 1 Flashcards
How psychological tests differ
- Content
- Response required
- Method of administration (group, individual)
- Use (Criterion or norm)
- Timing (power vs speed)
- Meaning of indicators (reflective vs formative)
Reflective measurment
Construct is assumed to cause differences in test scores. Item responses are indicators of the construct.
Items are necessarily correlated.
Formative measurement
Item responses define the construct. Items not necessarily correlated
Challenges in Psychological research
- Complexity
- Reactivity
- Observer bias
- Composite scores
- Sensitivity
- Awareness about the test
Factor analysis
A statistical method to study dimensionality of a test
Steps in exploratory factor analysis
- Correlation matrix
- Check Eigenvalues
- Select number of factors
- Interpret factors
Orthogonal rotation results in
Rotated factor matrix
Factors are uncorrelated
Oblique rotation results in
Pattern matrix
Structure matrix
Factors are correlated
Pattern matrix
Factor loadings controlled for the correlation between the factors (best interpretable)
Structure matrix
Correlation between item and factor (more difficult to interpret)
Characteristics of covariance
- It doesn’t convey magnitude
- It is a measure of linear association
- Gives info on the direction of the association
When to use normalization transformation to obtain normalized scores
When raw test scores are not normally distributed
Assumption for linking z scores to percentile rank
Individual differences have to be normally distributed
Why eigenvalue-greater-than-one rule is bad
Likely overestimating the number of factors
For what test is it meaningful to compute total test score
Multi-dimensional with higher order factor
