Search

The Even Start program, designed to demonstrate ways of implementing and delivering family literacy services, has had strong evaluation requirements since its inception. An analysis was conducted to determine the types of evaluations programs had conducted and their findings concerning Even Start outcomes. A convenience sample of 4 state evaluations and 118 local evaluations from 19 states were used for the analysis. Findings indicated that programs rarely used data for program management and improvement. Almost all evaluations reported some information on project implementation, including descriptions of project structure and activities, recruiting and retaining families, and working with collaborative agencies. More than three-quarters of the evaluations contained information about program outcomes, although none incorporated the use of control or equivalent comparison families. Twenty-four evaluations were selected for detailed examination of outcomes. These evaluations showed consistent positive gains on measures of child and adult development. Half the reports gauged the size of outcomes by making comparisons to national data, to evaluations of other early childhood or adult education programs, or to normal development. Because local studies did not provide much useful information to local project directors, State Coordinators, or to the Department of Education, several recommendations were made, including changes in Even Start legislation language, provision of guidance regarding the amount of funding needed for conducting evaluations focusing on program improvement, and the establishment of a community of local evaluators. (Includes a list of principles for the continuous improvement approach.) (KB)

Follow Through is a longitudinal program instituted by the U.S. Office of Education in which a variety of curriculum models have been implemented in a number of locations nationwide. Because services are provided from kindergarten through third grade, many children enter and leave the program at points not coterminous with the treatment itself. The purpose of this study was to assess the effects of attrition on the national Follow Through evaluation. Relationships between child characteristics such as ethnicity, sex, family income, and mother's education and attrition group membership were investigated to determine whether the findings of the major analysis should be qualified. (Author)

One specific correction model suggested by Cohen and Cohen (1975) is applied to data collected in the evaluation of a large-scale quasi-experimental program (Project Follow Through), and the effects of different assumptions about test reliability on the analysis results and on the conclusions of the evaluators are examined. The study determines whether the application of reliability or "true score" corrections alters the results obtained via an analysis employing uncorrected covariates in such a fashion as to appreciably change the policy-oriented conclusions of an evaluator. The data on which this paper is based were collected for the 1976 Follow Through evaluation and include measures on a total of over 5,000 children who began the program in kindergarten (Fall 1971) and completed it in third-grade (Spring 1975). Results indicate that application of true-score corrections using three separate reliability estimates to covariates employed on analysis of covariance did not change the conclusions of the Follow Through evaluators. (RC)

Describes two-generation programs and how they differ from earlier single-focus approaches to serve children and families. Results from six programs are reviewed and indicate mixed and modest success in promoting the development of children and improving the parenting skills and economic self-sufficiency of parents. Recommendations are presented for program improvement and future research. (GR)