Measurement Theory & Assessment II
Measurement Theory & Assessment II
Lectures
Week 1
Week 2
Week 3
Week 4
Week 5
Week 6
Tutorials
Tutorial 1
Tutorial 2
Tutorial 3
Tutorial 4
Tutorial 5
, Lectures
Week 1
Learning goals:
Latent variable, what complications occur, latent vs non-latent
Psychometrics, origins, roles, theory, causality, statistics?
Psychological test & types
Path diagrams
Reflective & informative measurement
How much information value a given variable has (in terms of the identity, order, quantity, and
absolute zero properties)
Different scales of measurement & implications
Exam 40 Questions
- No calculator
- No formulas
- 4-6 questions from the book
- Week 6 = practice exam
Assignment SPSS analysis
- Factor analysis & regression
- Concept version, feedback then final
Grading 40% assignment, 60% Exam
- Both need to be passing
Lecture 1 - Canvas Powerpoint Link
Psychological measurement Assignment of numerals to psychological characteristics of individuals
Psychometrics Attributes and psychological tests, how you measure
Latent variables (latent traits) = psychological variables
- Not directly observable
- We need to base it on something we can observe
We need:
1) Psychological theory
2) Causality
3) Statistics
4) An explicit graphical representation (path diagram)
Path diagrams = representation of a latent variable and their observed variables
- Latent variable: Showed in circle/ellipse
- Observable items:
Boxes/squares
- Errors: circle/ellipse
- Measurement error is
always unobserved
- Arrows: represent a
directional cause
, Recap Standard deviation: variance
- Negatively skewed
- Positively skewed:
- Normal distribution
Regression We need statistics to make the link between latent and observed - regression
model + correlation
Linear regression: the dependent variable is continuously distributed (WB1)
Logistic regression: the dependent variable is binary or dichotomous (WB2)
Association between Pearson's/Correlation coefficient: the linear relationship between two
variables variables (-1 to 1)
- Negative correlation: -
- Positive correlation: +
Expresses the consistency of the individual differences across two variables
Correlation doesn't tell us everything!
- Always look at your data before doing analysis
Importance:
- Items which measure the same thing (latent variable) should be
correlated
- Ex. someone who is extraverted will answer yes on both
questions (I like to talk to people, I like to party), shows that
both are measuring the same thing
Correlation matrix
Psychological Theory Tells you what the relevant observable variables are
- Ex. personality = big five
Numerals & measurement Levels:
levels - Nominal
- Ordinal
- Interval
- Ratio
Properties:
- Identity
- Order
- Quantity
Measurement Theory & Assessment II
Lectures
Week 1
Week 2
Week 3
Week 4
Week 5
Week 6
Tutorials
Tutorial 1
Tutorial 2
Tutorial 3
Tutorial 4
Tutorial 5
, Lectures
Week 1
Learning goals:
Latent variable, what complications occur, latent vs non-latent
Psychometrics, origins, roles, theory, causality, statistics?
Psychological test & types
Path diagrams
Reflective & informative measurement
How much information value a given variable has (in terms of the identity, order, quantity, and
absolute zero properties)
Different scales of measurement & implications
Exam 40 Questions
- No calculator
- No formulas
- 4-6 questions from the book
- Week 6 = practice exam
Assignment SPSS analysis
- Factor analysis & regression
- Concept version, feedback then final
Grading 40% assignment, 60% Exam
- Both need to be passing
Lecture 1 - Canvas Powerpoint Link
Psychological measurement Assignment of numerals to psychological characteristics of individuals
Psychometrics Attributes and psychological tests, how you measure
Latent variables (latent traits) = psychological variables
- Not directly observable
- We need to base it on something we can observe
We need:
1) Psychological theory
2) Causality
3) Statistics
4) An explicit graphical representation (path diagram)
Path diagrams = representation of a latent variable and their observed variables
- Latent variable: Showed in circle/ellipse
- Observable items:
Boxes/squares
- Errors: circle/ellipse
- Measurement error is
always unobserved
- Arrows: represent a
directional cause
, Recap Standard deviation: variance
- Negatively skewed
- Positively skewed:
- Normal distribution
Regression We need statistics to make the link between latent and observed - regression
model + correlation
Linear regression: the dependent variable is continuously distributed (WB1)
Logistic regression: the dependent variable is binary or dichotomous (WB2)
Association between Pearson's/Correlation coefficient: the linear relationship between two
variables variables (-1 to 1)
- Negative correlation: -
- Positive correlation: +
Expresses the consistency of the individual differences across two variables
Correlation doesn't tell us everything!
- Always look at your data before doing analysis
Importance:
- Items which measure the same thing (latent variable) should be
correlated
- Ex. someone who is extraverted will answer yes on both
questions (I like to talk to people, I like to party), shows that
both are measuring the same thing
Correlation matrix
Psychological Theory Tells you what the relevant observable variables are
- Ex. personality = big five
Numerals & measurement Levels:
levels - Nominal
- Ordinal
- Interval
- Ratio
Properties:
- Identity
- Order
- Quantity
