應用概化理論估計基本護理學技術測驗變異組成份之研究

Applications of Generalizability Theory to Estimate the Variance Components in Basic Nursing Technical Tests

本研究目的旨在應用概化理論估計基本護理學技術測驗各構面的變異來源。本研究以台灣地區某科技大學護理科五專二年級學生為研究對象，隨機抽取55位同意錄影的學生作為本研究之研究樣本。採用概化理論的two-facet crossed design，即：p × i × r設計，測量目標為學生，測量的構面為試題和評分者。以GENOVA軟體進行資料處理，第一階段先進行G研究分析，第二階段再進行D研究分析。本研究G研究的結果顯示：最重要的變異成份是「受試者」，其次是受試者（p）× 評分者（r）交互作用，第三是受試者（p）× 試題（i）× 評分者（r）。這三者佔總變異成份的97.77%，其餘四種變異來源所佔的變異成份很少。本研究D研究的結果顯示：增加評分者人數有助提高基本護理學技術測驗「無菌技術」實作評量的測驗信度。期望本研究結果能提供護理教師進行基本護理學技術測驗之參考。

The purpose of this study was to estimate the variance components in basic nursing technical tests by applying the generalizability theory. The 55 subjects in this study were selected at random from among second year nursing students enrolled at a 5-year junior college. A two-facet crossed design (i.e. p ´ i ´ r design) in the Generalizability Theory was used for this study, with students treated as the objects of measurement and items and raters treated as measurement facets. GENOVA software was used as the primary data analysis tool, followed by two additional steps, i.e., the G and D studies. G study findings were as follows: the examinee was the main source of variation was the examinee, with p ´ r representing a secondary source and the interaction of p ´ i ´ r representing a third. Nearly all variance (97.77%) was represented by these three factors. Another four sources of variation identified had only insignificant effect. The finding in the D study was that increasing rater numbers can promote reliability in basic nursing technical performance assessments. These findings are expected to provide a reference for teachers charged to develop or revise basic nursing technical tests.