Discontinuities in the development of mother-infant relationships:
applying a two-piece linear model to a study of the relationship between outcomes and underlying processes

M.V. Hernández-Lloreda1, F. Colmenares2 and R. Martínez-Arias1

1Departamento de Psicobiología, Universidad Complutense de Madrid, Madrid, Spain
2Departamento de Psicología y Educación, Universidad Camilo José Cela, Madrid, Spain

 

It is a well-established principle that behavior and social relationships change over time, and that these changes tend to be more dramatic during the early stages of ontogeny. A major challenge for developmental behavioral scientists is to find mathematical functions that capture the nature of these changes, and that allow tests of hypotheses regarding the possible effect of a number of variables:

The application of multi-level hierarchical models to study patterns of change in mother-infant relationships has shown that they can deal successfully with the first and second issues. Here, we use a two-piece linear model to explore its utility in analysing the third issue. Behavioral data from 23 mother-infant pairs of hamadryas baboons at Madrid Zoo were collected for 25 two-week periods over the first year of life, using focal animal and time sampling methods. We present results from the analysis of one of the 28 behavioral units considered: the percentage of time that the infant spent within 50 cm of its mother. We explore the effect of three variables known to influence spatial relationships between infants and mothers in primates: the infant's sex, and the mother's age and reproductive experience.

Our findings demonstrate a curvilinear relationship between time spent together and age, i.e. the multi-level hierarchical model was able to detect a 'discontinuity' in the behavioral outcome. Application of the two-piece linear model was then able to show that the association between the mother's age and the developmental course of this spatial unit changed when the two periods (i.e. before and after the occurrence of the discontinuity) were analysed separately. This suggests that the observed discontinuity in the behavioral outcome may be driven by a discontinuity in underlying processes linked to the mother's age. We conclude that multi-level hierarchical models in general, and piecewise linear models in particular, can be very helpful in unravelling the principles of behavioral development.


Paper presented at Measuring Behavior 2002 , 4th International Conference on Methods and Techniques in Behavioral Research, 27-30 August 2002, Amsterdam, The Netherlands

© 2002 Noldus Information Technology bv