Jill and Peter de Villiers are co-PIs of a large NIH grant to explore language and Theory of Mind outcomes in preschool children growing up in poverty. The grant is part of a much larger program grant, the focus of which is on preschool teacher and curriculum development to improve school readiness of children at high risk for school failure. Several intertwined research projects are designed to follow the children’s progress and outcomes. The participants are two samples of high-risk, primarily Hispanic and African-American children aged 2-3-years old and 4-5 years old, their parents, and their teachers. The project allows the precise assessment of the children over the year - at the beginning, middle, and end of the interventions and 1 year later (as well as 2 years later for the older group).
The program project has several sites:
University of Texas Medical Center, Houston TX (PI: Susan Landry)
Florida State University, Tallahassee, FL (PI: Chris Lonigan)
University of Berkeley, CA (PIs: Alice Klein and Prentice Starkey)
University of Arizona, Tucson, AZ (PI: Nancy Eisenberg)
University of Guelph, Canada (PI: Marcia Barnes)
Smith College (PIs: Jill and Peter de Villiers)
The Florida and Houston sites are the main centers of activity for training the teachers, monitoring the curriculum, and testing the children, and the other University sites are carrying out specific coding and analyses.
In our part of the grant here at Smith, we will provide an in-depth analysis of the language and theory of mind (ToM) skills developed in the children who participate in the larger intervention program project.
We will use the data gathered during this project to test competing models of the interrelationships among the measures, allowing significant theoretical progress in the fields of language and cognitive development quite independent of any outcome variations due to curricula. In turn, this study will also complement related projects on the same children dealing with social-emotional development and mathematics readiness.
Measures of meta-representation, executive function (specifically working memory and inhibitory control), early ToM and False Belief reasoning tasks (standard and non-standard) are given to the children together with the language tasks in a longitudinal design, and models of their interrelationships will be tested at several time points. These empirical data will contribute significantly to the theoretical debates about the precursors of mature ToM.
Finally, in addition to the enriched language measures, ToM outcomes will be used to predict later social functioning and school readiness in the groups of children, relations that have been theoretically predicted but not fully explored in practice.
A great deal of activity is involved in transcribing videotapes sent to us from the testing sites. We are transcribing:
a) the children telling narratives
b) the children and their mothers in a play session
c) the mothers describing their children
d) the teachers in ordinary classroom activities at each site
For each group, we have designed a computer database to allow coding along multiple dimensions. Here is a typical hard-at-work transcriber!
In addition, we need to make materials for testing all kinds of language and theory of mind skills. We have burned out two laminators making pictures that will survive hundreds of fingerprints and last four years. We also have had to use creativity in making novel objects for children to name. We go to a store with the aim of finding interesting things that will capture two year olds’ attention, that do not have an obvious name, cannot be swallowed or otherwise cause harm, and cost less than a buck each! Here is a group picture from our last “widget sweatshop” in 2006.
Since many of the children we are testing are not speakers of the “Standard English” that tests are designed around, we are using innovative new materials from both the African American English assessment we helped to create (the DELV), and a new test under development for testing bilingual Spanish-English speakers, the BESA (developed by Elizabeth Pena, Aqiles Iglesias and Lisa Bedore). We have employed bilingual Spanish speakers to help with transcription, and brainstormed about sensitive measures of mother-child interaction for all these cultural groups. The children are also receiving lots of standard testing, so we want to be sure that we also take their cultural backgrounds into account.