背景 : 護理人力在醫療機構中是最重要的照護主力，現今面臨護理人力短缺，而護理應屆畢業生為潛在的人力資源，其工作選擇機構為一重要議題。 目的 : 旨在建立評估護理科系應屆畢業生工作機構選擇之量表。 方法 : 採方便取樣台灣南部護理科系應屆畢業生，分別進行非結構式問卷調查、結構問卷預試及正式施測，利用 SPSS13.0進行獨立樣本 t檢定、信度與因素分析，及 LISREL8.72版以結構方程模式進行驗證性因素分析，驗證量表之建構效度。 結果 : 以 100份非結構式問卷調查結果，編制結構式問卷 36題，經專家內容效度檢定後成為 29題。預試收案 270人，以樣本項目分析後刪除 3題；探索性因素分析萃取五個因素：工作環境、員工與機構發展、重要關係、機構地理特性、排班條件，總解釋變異量為 63.794%。全量表之 Cronbach’s α為 .936。正式施測收案 295人，構面題項與量表總分皆顯著相關；以驗證性因素分析，因素負荷量介於 .574~.787；全量表之內在一致性為 .921，五個因素信度為 .833、.822、.682、.711及 .623，變項的組成信度為 .675~.882。 結論／實務應用 : 本量表共 26題均合適且具建構效度，為簡短且易於施測之工具，可運用於相關研究。

Background: While nurses predominate in the healthcare institution workforce, Taiwan continues to face a chronic nursing shortage. Nursing program graduation candidates are a critical human resource and their selection of an institution appropriate to their career interests and needs is an important issue. Purpose: This research established an Institutional Selections Scale of Nursing Graduation Candidates (ISSNGC). Methods: We selected subjects from among graduation candidates enrolled at nursing programs in Southern Taiwan. Researchers used unstructured interviews and a questionnaire survey to collect data. A pilot survey was initially conducted prior to running the formal survey. SPSS13.0 and LISREL8.72 used structural equation modeling (SEM) to analyze data. Results: Researchers initially interviewed 100 subjects using an unstructured questionnaire that had been reviewed by experts for validity. Refinements further reduced the number of questions to twenty-nine. A 270-person sample was targeted for pilot test and sample analysis. The result eliminated a further three questions. Exploratory factor analysis was then performed on data. Analysis yielded a five-factor solution that explained 63.749% of variance. The five factors were labeled “Working Environment,” “Employee and Institution Development,” “Important Relationships,” “Institutions’ Geographical Characteristics” and “Staffing.” The Cronbach’s α for the ISSNGC was .936. Researchers then recruited a total of 295 people for the formal test. Correlation coefficients for items and the overall ISSNGC score were significant. Confirmatory factor analysis identified five factors. Each absolute factor loading value fell between .574 and .787. Internal consistency reliability for the five factors were .833, .822, .682, .711 and .623, respectively, and .921 for overall ISSNGC. Composite reliability fell between .675 and .882. Conclusions / Implications for Practice: ISSNGC offers appropriate construct validity in a simple, yet efficient measurement tool. It is recommended for use in related research.