The assessment procedure to conduct a PIR - GAS rating on a 10 - minute interaction sample seems to allow a reliable, but not
necessarily valid information about the mother - child - relationship quality.
With regard to the sample size required for
valid application of this modeling approach, we are hesitant to speak of sufficient sample sizes, since a large number of observations with little score variation over time (as in our second empirical application) does not
necessarily provide richer
information than a smaller sample with more fluctuations.