Reliability Estimation for Aggregated Data: Applications for Organizational Research
Abstract
In order to study organizations it is important to be able to measure organizational functioning with a minimum of error. The report that follows provides the statistical tools necessary to measure the extent of error that exists in survey data, and organizational record data. Traditional methods of measuring error are either inappropriate or incomplete when applied to organizational groups, necessitating the statistical development given here. Appropriate methods of measuring error are particularly important when organizational change is being studied. In this case, the same variables are measured at more than one point in time. The investigator wants to identify real organizational change. However, real change cannot be separated from changes in measurement error, unless separate estimates of measurement error are available at each point in time. This paper tells how to get separate error estimates so that real organizational change can be studied. A statistical technique for estimating the level of the hierarchy that actually controls the subject matter at hand is provided. This measure can be used as a guide for selecting groups at appropriate levels of hierarchy for study. These statistical techniques provide improved procedures for studying the operation of the Army and other organizations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1981
- Accession Number
- ADA129740
Entities
People
- Roland J. Hart
- Stephen C. Bradshaw
Organizations
- U.S. Army Research Institute for the Behavioral and Social Sciences