Generalizabiity and dependability of brief behavior rating scales for social skills

While there are appropriate tools to progress monitor academics, there is no universally accepted tool to progress monitor social behavior. The current study proposed the development of a series of brief behavior rating scales to correspond to important social skills domains on the Social Skills Rating Scale (Gresham & Elliott, 2008) and to evaluate the resulting psychometric features through generalizability theory. Data was collected in a preschool classroom in a 6 persons by 2 rater by 6-7 items by occasions mixed model design. Data was analyzed series of generalizability and decision studies to investigate sources of variability and to determine the assessment conditions required to make reliable decisions. Results indicated that large proportions of the total variance in these scales were attributed to rater-related effects. This affected generalizability and dependability coefficients to the extent that reliable decisions could not be made with the current scale using feasible assessment conditions. Furthermore, current results did not support the abbreviation of the current scales. These finding indicates the need to further understand and control for unwanted variability between raters. Implications for the assessment of social behavior and suggestions for scale development are reviewed.

Name thesis:degree_name

Doctor of Philosophy (PhD)

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Psychology

Grantor

Psychology

Year dc:date.available

2013

Author dc:creator

• Minor, Lisa L

• social skills assessment
• generalizability theory
• reliability

Statement dc:rights

• unrestricted
• Release the entire work immediately for access worldwide.

Identifier

etd-10162013-234347
https://repository.lsu.edu/gradschool_dissertations/281

OAI identifier oai:identifier

oai:repository.lsu.edu:gradschool_dissertations-1280