Information for Tribal Start Applicants on Selecting Adults and Early Childhood Assessments
Tribal Even Start grantees are required to report annually to the Secretary regarding the progress made towards addressing the purposes of the program. Specifically, grantees will be required to report on the following GPRA measures:
- Measure 1: The percentage of Tribal Even Start adults who do not have limited English Proficiency (LEP) who achieve significant learning gains in reading.
- Measure 2: The percentage of Tribal Even Start adults with limited English Proficiency (LEP) who achieve significant learning gains in reading/English Language acquisition.
- Measure 3: The percentage of Tribal Even Start adults with a high school completion goal who earn a high school diploma.
- Measure 4: The percentage of Tribal Even Start adults with the goal of General Equivalency Diploma (GED) attainment who earn a GED.
- Measure 5: The percentage of preschool-aged children participating in Tribal Even Start who achieve significant gains in oral language skills.
- Measures 6: The average number of letters Tribal Even Start preschool-aged children are able to identify.
- Measure 7: The percentage of preschool-aged children participating in Tribal Even Start who demonstrate age-appropriate oral language skills.
While grantees may select their own assessment instruments for use in reporting on these measures, this document provides some basic guidelines for selecting quality assessments.
1. For measures 1,2,5,6, and 7, the grantee should be able to demonstrate that the oral language assessment selected is:
- a. Valid for the subject and age range of students being tested. Validity is the extent to which the test measures what it was supposed to (e.g., reading aptitude, oral language skills). There are several types of validity including content, construct, and predictive validity. Evidence of validity is often available from the publisher of the assessment instrument.
- b. Reliable with regard to the consistency and repeatability of measurement. Several types of reliability are routinely established, including the internal consistency of an instrument (how well different items on an instrument measure the same construct) and test/retest reliability (the consistency of measurement at two different points in time). Evidence of reliability is often available from the publisher of the assessment instrument.
2. For assessments used to determine if an adult or child made a significant gain (for measures 1,2, and 5), the same assessment should be used for pre- and post-tests. For assessments that have different versions of the assessment (e.g., for different age ranges), the grantee should ensure data received from the pre-assessment can be compared to data received from the post-assessment to calculate a change in score. The test publisher can inform the grantee whether pre- and post-assessment data from different versions can be compared.
3. For assessments used to determine whether a child is demonstrating age-appropriate oral language skills (for measure 7), the assessment should have information about normative statistics for the instrument. This information is commonly referred to as “population norms” or “normed-referencing data” and is provided by the instrument’s author or publisher. Since young children’s skills can develop rapidly because of maturation, we recommend that assessments used for measure 7 have normed-referencing data available in 6-month or smaller age intervals. For example, an assessment provides the mean and standard deviation of the norm-referencing data for ages 48 to 53 months, 54 to 59 months, etc.
4. In addition to the characteristics described above, the evaluators for the National Evaluation of Early Reading First described some additional factors they considered when selecting assessments. This list of factors may be helpful to applicants as they select their assessment instruments.
- Time required to administer the instruments
- Training required for staff to administer the instruments
- Qualifications that examiners needed so that appropriate and adequate staff were trained and available
- Sensitivity of the measures to change as a result of the intervention
- Appropriateness of the measure for a diverse population including racial and ethnic minorities, language minorities, and economically disadvantaged children
- Costs of the measures for the sample sizes
- Comparability of the measures to other national evaluation studies (especially other current early literacy intervention studies)
- Psychometric qualities of the measures under consideration, including adequate reliability and validity, with minimal floor or ceiling effects for low-income preschool children
- Availability of a Spanish-language version of assessment
Russell Jackson, McCoy, Ann, Pistorino, Carol, Wilkinson, Anna, Burghardt, John, Clark, Melissa, Ross Christine, Schochet, Peter and Swank, Paul. National Evaluation of Early Reading First: Final Report, U.S. Department of Education, Institute of Education Sciences, Washington, DC: U.S. Government Printing Office, 2007, p. 161.