ERIC Number: ED249275
Record Type: RIE
Publication Date: 1983-Aug
Reference Count: 0
Reliability and Validity of 10 Different Standard Setting Procedures.
Halpin, Glennelle; Halpin, Gerald
Research indicating that different cut-off points result from the use of different standard-setting techniques leaves decision makers with a disturbing dilemma: Which standard-setting method is best? This investigation of the reliability and validity of 10 different standard-setting approaches was designed to provide information that might help answer that question. The 10 procedures for setting a standard on the Missouri College English Test included: a normative method (33rd percentile), the chance/ideal mean approach, the Ebel method, the Nedelsky method, the Angoff method, and five methods comparing different subsets of practicing teachers. Phi coefficients correlating pass/fail decisions for all two-method combinations of 10 standard-setting procedures ranged from .16 to 1.00 indicating greater consistency or agreement between some methods (e.g., practitioners--borderline group) than others (e.g., chance/ideal mean--masters). Phi coefficients between pass/fail with the 10 standard-setting methods and pass/fail on an external criterion ranged from .20 to .40 indicating greater validity for some methods (e.g., practitioners and borderline group) than for others (e.g., non-masters). (Author/BW)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Angoff Methods; Ebel Method; Nedelsky Method
Note: Paper presented at the Annual Meeting of the American Psychological Association (91st, Anaheim, CA, August 26-30, 1983).