Simple Regression with an Extended Error Term Flashcards
What issue does this type of simple regression improve?
The causal issue as it is based on longitudinal data.
What is the fixed intercept?
Y-intercept (Bo).
What is the fixed slope?
The gradient (B1).
What are the random intercepts?
The intercept of each individual person.
What is roi?
The deviation from the fixed intercept to the random intercepts.
What are the random slopes?
The individual slope of each person.
What is r1?
The deviation from the fixed slope to the random slopes.
How is the error measured?
It is the deviation from the observed data point to the individual slope at each specific point in time.
Why is this new model statistically better than GCM?
Error term is smaller - can trust results more.
Where can you find the fixed effects in SPSS?
Estimates of fixed effects.
intercept - fixed intercept.
predictor - fixed slope (this is what you should report with the p-value).
Where can you find the random effects?
Estimates of covariance parameters.
Residual - error term. intercept - Ro.
predictor - R1.
What is MLM?
It is the modern way of studying how outcomes at the individual level can be seen as the results of the interplay between individual and contextual factors (nested data).
What are the four levels of nested data?
Individuals nested in a context.
Changes over time - longitudinal data.
Processes within persons over time - intensive longitudinal data.
Multi-modal real-time processes.
These are in different time scales.
Give some examples of people nested in a context.
Patients nested in therapists.
What is the difference between people and processes?
Associations can change depending on whether its within a singular person or within many people.
What is the difference between traits and processes?
Traits - do not change over time.
Processes - implies some kind of within-person variability over time/situation.
What does Hilpert believe about most theories?
That they study within-person processes.
Explain how multi-modal real-time processes can be studied?
Observational coding - but this produces spares data.
Need new tools which detect and extract behaviour automatically.
Allows for more complicated analysis (e.g. DSM).
What is growth curve modelling?
Concerned with how psychological processes unfold over time.
What is the minimum amount of time points necessary?
3.
What about assumption checks?
Very complicated to conduct - just need to know the idea of it.
MLM can often solve the problem of interdependence.
What is a limitation of GCM against intensive longitudinal data?
Focuses on average change across many people.
Can’t understands processes - changes in a person over time.