頁籤選單縮合
題 名 | 量性護理研究論文之分析: 研究設計、信效度、推論統計方法=Examination of Quantitative Nursing Research: Study Design, Validity and Reliability, Inferential Statistics |
---|---|
作 者 | 史麗珠; | 書刊名 | 長庚護理 |
卷 期 | 9:2=22 1998.06[民87.06] |
頁 次 | 頁23-31 |
分類號 | 419.63 |
關鍵詞 | 量性研究; 研究設計; 信效度; 推論統計; Quantitative research; Study design; Validity and reliability; Statistics; |
語 文 | 中文(Chinese) |
中文摘要 | 雖然近年國內護理研究質數不斷提升,但尚有很多可改善的地方。本文即針對國 內護理研究期刊,刊登量性研究論文之研究設計、抽樣方法、研究工具之信效度、推論統計 方法,分析其常用性,適合性,作為提升護理研究素質之參考。 結果自86年的「護理研究」、「長庚護理」、「榮總護理」摘錄67篇量性研究論文。以 敘述性或分析性研究論文最多(70.1%),類實驗研究次之(26.9%),建立量表之信 效度量少(11.9%)。約一半的論文採用立意或方便採樣,使論文結果之推廣性受到限制。這 些論文共採用118個量表,參考文獻後自行研製量表有61個(51.7%),翻譯國外量表有25個(2 1.2%),使用現成量表有32個(27.1%)。大部份使用的量表都有作信效度的評估,效度評估 以專家效度為主,信度評估則以Cronbach's �`為主,顯示其他信效度評估仍有待加強。又 發現有一半的自行研製或翻譯自國外之量表都沒有將量表全部內容列於論文中,其他研究者 難以重覆相同研究,研究成果也較難落實。在統計推論之應用方面,具顯著統計意義的單因 子變異數結果沒有作多重比較,沒有明確說明使用哪種統計方法計算相關係數或線性複迴歸 ,沒有檢定迴歸之統計假設前題,顯著水平與P值之混淆等,都顯示部份護理研究者在統計 推論方法上的應用不夠嚴謹。 |
英文摘要 | The quality of nursing research have been improving in a great magnitu- de currently.However, we should continue our effort on other aspects in order to gain respects from non-nursing researchers. The aim of this study was to evalua- te the appropriateness of study design, sampling methods, validity and reliabil- ity of instruments, and inferentila statistics used in quantitative nursing res- earch. One-year publications in three domestic nursing journals (including Nursing Research, Chang Gung Nursing,VGH Nursing) (n=67) were evaluated.Descriptive or analytical study was most common (70.1%),followed by quasi-experiments (26.9%), and assessment of validity and reliablility of a instrument (11.9%).About half of the studies used purposive or convenient method to obtain the study sample, limiting its ability in generalization of the results. A total of 118 instruments used in htese papers, 61 were developed by the researchers, 25 were translated from foreign questionnaires, and 32 used existing Chinese questionnaires. Expert validity and Cronbach's �` was primarily used for evaluation of validity and reliability, respectively, suggessting other assessments in validity and reliability should be encouraged as well.About half of the instruments from developed by the researchers or translated foreign questionnaires did not show the whole content, creating some difficulty for other researchers to reproduce similar studies. Problems such as multiple comparisons not used or not mentioned in a significant one-way ANOVA, not specified which methods for correlation coefficients or for multiple linear regression, not examined the statistical assumptions in multiple linear regression, confusion between significant level and p-values, indicated some of nursing researchers are not familiar with statistical methods. |
本系統中英文摘要資訊取自各篇刊載內容。