Selecting and Using Early Childhood Rating Scales

Abstract

The purpose of this article is to help rating scale users organize their thinking about why and how rating scales are to be used and to provide them with information that will be useful in evaluating the adequacy of available rating scales. The first part of the article discusses issues related to establishing a rationale for the use of rating scales, pointing out questions that should be asked regarding purposes for and means of implementing scale use. The second part discusses issues related to selecting a rating scale to meet user needs, such as user friendliness, scale format, score interpretation, reliability, and validity. The section on reliability contains an expanded discussion of how to evaluate the reliability of a scale and the factors that can affect reliability. The validity section focuses discussion on two questions: Does the scale measure the trait validly? Can the scale be validly used to serve its intended purposes? Various methods for establishing the validity of a rating scale are discussed in relation to these two questions. Special attention is given to discussing the effects that rating scale score distributions can have on a scale's validity for specific purposes.

The Validity of Teacher Ratings of Adolescents' Social Skills

“…But since the validity of the SSRS teacher form has not been examined in the Norwegian context before, and even though earlier studies have supported its validity , 1990Elliott & Gresham, 1989;Demaray et al, 1995), a reassessment was considered necessary when the rating system was translated and slightly modified for use in this study. For this purpose, a school cohort of 395 adolescents in a Norwegian municipality was assessed in seventh and ninth grade with the SSRS.…”

“…Internal consistency measured with Cronbach's alpha communicates to what extent the items measure the same construct with coefficients higher than r ϭ 0.80 considered satisfactory and coefficients exceeding r ϭ 0.90 as excellent. Test-retest reliability as a measure of stability in assessment over time is probably the most important reliability measure in behaviour ratings (McCloskey, 1990), where test-retest correlations larger than r ϭ 0.80 signal high stability, and even coefficients higher than r ϭ 0.70 are acceptable for time intervals of 6 months or more (Elliott et al, 1993). Inter-rater reliability expresses correspondence in the ratings from different raters of the same individual, and since different raters contribute unique views of the subject being rated, the ratings may vary substantially.…”

Documenting changes in communication behaviours using a Structured Observation System

“…Types of assistance or support included verbal models, repetition of cues or instructions and physical assistance. The numerical rating scale used was developed according to the guidelines suggested by McCloskey (1990) and included the following descriptions: 0 ϭ Behaviour not observed. When a particular behaviour was never observed, even with maximum support from an adult, a rating score of '0' was assigned.…”

Development of a performance-rating scale for a nutrition knowledge test developed for adolescents

“…Performancerating scales can be criterion-referenced, content-referenced or norm-referenced (6) . A criterion-referenced performancerating scale rates a person's performance in relation to mastery levels (6) , which is the testee's ability to perform a given set of competencies independently of other testtakers (7) .…”