Molenaar - reading questions Flashcards
(7 cards)
Ergodic process
when model at cross-sectional level (1 timepoint, many people - interindividual information) is the same as model that holds for individuals when they were measured over time (intraindividual information)
How do you get from phase space to behaviour space?
add time as dimension
Explain difference between random and deterministic trajectory
In a random trajectory your position in the behavior space at timepoint t+1 is not fully determined by
your position at time point t.
In deterministic trajectory you can fully predict from time t where somebody will be in time t + 1
What stationarity means?
Stationarity means that the parameters in the model do not change over time.
Note that this does not mean that the variables do not change (your scores can vary from time to time), but parameters such as the means, variances, covariances, factor loadings, regression coefficients etc cannot change over time.
Explain Molenaar simulation study and what was derived from the results
For each person you can estimate a factor model based on multiple timepoints from their trajectory: IAV (intra individual) model.
Molenaar shows that when you simulate data in which each person has their own IAV model, then taking a cross section of the data and estimating an IEV (interindividual) model, this will result in a well fitting factor model.
(with large samples, one-or two-factor models will fit the data). This means that IEV factor analysis seems to be insensitive to substantial heterogeneity! It thus could
be the case that a one or two-factor model fits well in cross sectional data, even though the data is
generated from a population where each person has their own different model over time!
What happened in personality study?
IEV factor analysis supported 5 factor model. However, IAV factor analyses over time showed different factor model for each person (different factor loadings, different number of factors)
