Comparison of Nomothetic Versus Idiographic-Oriented Methods for Making Predictions About Distal Outcomes From Time Series Data

Abstract

We illustrate the idiographic/nomothetic debate by comparing 3 approaches to using daily self-report data on affect for predicting relationship quality and breakup. The 3 approaches included (a) the first day in the series of daily data; (b) the mean and variability of the daily series; and (c) parameters from dynamic factor analysis, a statistical model that uses all measurement occasions to estimate the structure and dynamics of the data. Our results indicated that data from the first measurement occasion does not provide information about the couples' relationship quality or breakup 1 to 2 years later. The mean and variability of the time series, however, were more informative: females' average positive and negative affect across time was related to relationship quality, whereas males' variability in negative affect across time was predictive of breakup. The dynamic factor analysis, in turn, allowed us to extract information central to the dyadic dynamics. This information proved useful to predict relationship quality but not breakup. The importance of examining intraindividual variability and couple dynamics is highlighted.

Notes

¹We purposely use the term idiographic-oriented to indicate that although this approach is based on modeling the individual unit of analysis, certain assumptions (described in the Discussion section) about these individual units are made to arrive at group-level estimates. Thus, this third approach is not idiographic in the purest sense of the word.

²The interpretation of the first measurement occasion variables as having lower reliability is arguable (see Hertzog & Nesselroade, 1987) because in our case these variables are a measure of state affect, which might be highly variable across time (suggesting low test-retest reliability) but can have high internal consistency.

³ACF and PACF plots are available from Laura Castro-Schilo upon request.

⁴The parcels for positive affect were created in the following fashion: Parcel 1 = enthuse, interest, strong. Parcel 2 = excited, determined, attentive. Parcel 3 = proud, inspired, alert, active. For negative affect: Parcel 1 = afraid, irritable, hostile. Parcel 2 = distress, nervous, ashamed. Parcel 3 = upset, scared, guilty, jittery. The assignment of items across parcels was based on the domain-representative method for parcel construction put forth by Kishton and Widaman (1994) and entailed factor analyses at the group level.