are consistent with those aimed at reforming welfare in relation to early childhood programs.
At the program level, several participants requested that federal and state governments provide more funding for training of providers of care to infants and toddlers. In fact, many participants felt that significant support was needed to build a workforce in early childhood education who could translate research findings to the program level, act as technical assistance advisors, and design and implement research initiatives that inform the field, programs, parents, and policy makers.
In the words of Martha Zaslow, of Child Trends, Inc., several areas of differences—or dialectics—about performance measures emerged at the workshops. Concerns about the purpose of performance measures arose throughout the workshops—the tension between information for accountability and information to improve services. The former implies sanctioning—funds are withheld if performance measures are not met; the latter, some participants felt, reflect the drive to make programs better.
Which level is most appropriate for developing performance measures also was an issue for debate. Many participants stressed that the vision and goals must emerge at the local or program level. On the other hand, there is a need for measures and indicators that have broad geographic applicability.
Differences also emerged on who should supply information—parents, program providers, or the children themselves? Most notably, participants differed about whether to emphasize careful delineation of child outcomes and expectations in child care or to focus on delivery of programs and services. Throughout, participants were keenly aware of the difficulties of attributing child outcomes to child care given the diversity of settings children are in and all the other influences on their lives.
In sum, developing performance measures entails what many saw as an arduous process of consensus-building on all these issues and on the challenges of identifying and aggregating reliable data. These workshops provided a step in this necessarily long-term process, and much more remains to be done.