國內三份期刊的醫管論文分析：抽樣方法、信效度、統計方法

Examination of Health Care and Hospital Administration Publications in Three Domestic Journals: Sampling Method, Validity and Reliability and Statistical Method

李芳年(Fang-Niarn Lee) ； 盧育誠(Yu-Cheng Lu) ； 陳曉真(Sheau-Jen Chen) ； 盧佩玲(Pei-Ling Lu) ； 史麗珠(Lai-Chu See)

《中華公共衛生雜誌》 19卷3期 (2000/06) Pp. 231-238

https://doi.org/10.6288/CJPH2000-19-03-09

摘要

目標：本研究摘錄國內三份期刊近五年醫管論文使用之抽樣方法、研究工具信效度及統計分析方法，以瞭解其常用性、適切性。方法：選取83下半年至88上半年「醫院」、「中華衛誌」、「公共衛生」的醫管原著論文，由本文作者經訓練後作登錄並作描述性統計。結果：共收錄109篇。以描述性或分析性（80%）為大宗，類實驗研究（17%）次之，模擬性的論文最少（3%），描述性或分析性研究中有42.5%採用立意或方便取樣。7篇（8%）論文對已達顯著意義的單因子變異數分析沒有進行多重比較，或4篇（4.6%）沒有說明使用哪種多重比較方法；12篇論文使用相關係數論文中8篇沒有特別說明是使用哪種係數，5篇（10.4%）誤認為p值就是顯著水平，迴歸模式沒有進行殘差分析。自行研製的量表有近1／4沒有將量表內容全部刊登在論文內，研究工具之效度以專家效度為主，一半沒有做信度檢定。結論：近年醫管論文約一半的描述性或分析性研究仍採用立意或方便取樣，樣本代表性易受質疑，部份統計推論方法不夠嚴謹，研究工具之效度專家效度為主，一半沒有做信度檢定。

關鍵字

抽樣方法 ； 信效度 ； 推論統計

並列摘要

Obiectives: The aim- of this study was to evaluate the appropriateness of sampling methods, validity and reliability or instruments, and statistical methods used in domestic publications of health care and hospital administration. Methods: Five-years health care and hospital administration publications in “Hospital”, “Chinese Journal of Public health”, “Public Health” (n=109) were evaluated. Abstractors were trained and descriptive statistic was used. Results: Descriptive or analytical study was most common (80%), followed by quasi-experiments (17%) and simulation study (3%). Purposive or convenient method was primarily used to obtain the study sample (42.5%), There were some problems such as failure to mention (4.6%) or non utilization (8%) of multiple comparisons in a significant one-way ANOV A, lack of specific methods used for correlation coefficients (66.7%), non-examination of statistical assumptions in multiple linear regression (100%), confusion between significant level and p-values (10.4%). About one-forth of the instruments developed by the researchers did not show the whole validity was primarily used for evaluation of validity. Reliability test for about half of the instruments were not carried out. Conclusions: Our evaluation on domestic health care and hospital management publications suggested that representative of the sample may be a problem in descriptive or analytical studies. Some statistical methods used were not stringent. Expert validity was primarily used in evaluation of validity of study. About hall’ or the instruments reliability tests were not performed.

並列關鍵字

sampliug method ； reliability alld validity ； statistical method.