Direct behavior rating: An evaluation of time-series interpretations as consequential validity

Theodore J. Christ, Peter M. Nelson, Ethan R. Van Norman, Sandra M. Chafouleas, T. Chris Riley-Tillman

Research output: Contribution to journalArticle

7 Scopus citations

Abstract

Direct Behavior Rating (DBR) is a repeatable and efficient method of behavior assessment that is used to document teacher perceptions of student behavior in the classroom. Time-series data can be graphically plotted and visually analyzed to evaluate patterns of behavior or intervention effects. This study evaluated the decision accuracy of novice raters who were presented with single-phase graphical plots of DBR data. Three behaviors (i.e., academically engaged, disruptive, and respectful) and three graphical trends (i.e., positive, no trend, and negative) were analyzed by 27 graduate and five undergraduate participants who had minimal visual analysis experience. All graphs were unique, with data points arranged to form one of three "true" trends. Raters correctly classified graphs with positive, no, and negative trends an average of 76, 98, and 67% of instances. The generalized linear mixed model was used to handle significance tests for the categorical data. Results indicate that accuracy was influenced by the trend direction, with the most accurate ratings in the no trend condition. Despite the significant effect for trend direction, the current study provides empirical evidence for accuracy of DBR trends and interpretations. Novice raters and visual analysts yielded accurate decisions regarding the trend of plotted data for student behavior.

Original languageEnglish (US)
Pages (from-to)157-170
Number of pages14
JournalSchool Psychology Quarterly
Volume29
Issue number2
DOIs
StatePublished - Jun 2014

All Science Journal Classification (ASJC) codes

  • Education
  • Developmental and Educational Psychology

Fingerprint Dive into the research topics of 'Direct behavior rating: An evaluation of time-series interpretations as consequential validity'. Together they form a unique fingerprint.

  • Cite this