Research Seed Grant | 2004-2005
A New Coding Scheme for the Five Minute Speech Sample in the Fragile Families Child and Wellbeing Study
by Jeanne Brooks-Gunn (Teachers College), Anne Martin (Teachers College), and Rebecca Ryan (Teachers College)
Current large-scale surveys focusing on children and families are limited in their measures of the parent-child relationship. Self-reports from parents and children about the nature of their relationship are subject to bias and social desirability. In addition, children may be too young to be interviewed. Some small-scale studies have successfully videotaped and coded parent-child interactions, but at great expense. We propose the development of a coding scheme for a possible alternative mode of data collection borrowed from clinical psychology—the Five Minute Speech Sample (FMSS). In the FMSS, parents are asked to describe their child and their relationship; responses are audio-taped and coded. The FMSS is currently being administered as part of the Fragile Families Child and Wellbeing Study, a national longitudinal study of unwed parents and their children. Compared to young children living in 2-parent families, those living with unwed parents or a single parent typically have less favorable developmental outcomes. We plan to examine whether this associations can be explained in part by mothers' emotions towards their children as expressed during the FMSS. Preliminary plans for coding the FMSS include scales capturing the valence (positivity, negativity, or both) of mothers' comments, as well as the kinds of attributes mothers use to describe their children. We expect that mothers' affection toward their children and the kinds of qualities they find most salient will vary according to family structure. Seed grant money will be used to develop the coding scales and, once they are finalized, prepare training materials for coders. We will seek external funding for coding all tapes and conducting data analysis after the scales are developed. If our FMSS measure proves successful, it should be adaptable for use in other large-scale family surveys.
