Beyond classical reliability: using generalizability theory to assess dependability.

Generalizability theory (GT), an approach to estimating measurement dependability, offers researchers the ability to assess comprehensively various sources of measurement error simultaneously. Application of GT identifies the conditions that contribute most substantially to error and offers remedies to enhance dependability. A review of GT is presented and a practical application discussed. Researchers who develop instruments or plan studies using observation or questionnaires will appreciate the advantages offered by GT.