查詢結果分析
相關文獻
- 施肥效應之統計分析與探討--以扁柏苗木施肥為例
- 試驗前苗木篩選對處理效應之影響--以臺灣扁柏苗圃氮施肥為例
- A-Optimal Designs for an Additive Quadratic Mixture Model
- 無參數試題反應理論的能力群組之模糊分割
- A Nonparametric Test for the Presence of Immunes in Type I Censored Data
- A Dislocation and Point Force Approach to the Boundary Element Method for Mixed Mode Crack Analysis of Plane Anisotropic Solids
- 常態混合模式的模糊分割之資料模擬研究
- 數量性狀基因座定位法在卜瓦松分布資料上之應用
- 數量性狀基因座定位法在二項分布資料上之應用
- The Application of Genetic Algorithm to Mixed Pixel Classification in Remote Sensing and MR Images
頁籤選單縮合
題 名 | 施肥效應之統計分析與探討--以扁柏苗木施肥為例=Nursery Nitrogen Fertilization of Taiwan Cypress--A Case Study on Statistical Analyses of Fertilization Experiments |
---|---|
作 者 | 施佩君; 關秉宗; | 書刊名 | 中華林學季刊 |
卷 期 | 31:4=123 1998.12[民87.12] |
頁 次 | 頁349-359 |
分類號 | 436.194 |
關鍵詞 | 重複觀測; 縱向研究; 混合模式; 一般線性模式; Repeated measures; Longitudinal study; Mixed model; General linear model; |
語 文 | 中文(Chinese) |
中文摘要 | 在育林研究上,常進行重複觀測的試驗,以瞭解處理措施的成效。然在重覆觀測的 試驗中,若不考慮時間之效應,則可能將處理效應與時間效應混淆,而不易評估處理真正的效 應;故正確的統計模式應考慮時間的因素,且分割時間的效應,以達到正確的結論。本試驗目 地在於討論四種統計模式對重複觀測試驗結論的影響,以選擇適當的統計分析方法獲得最多的 資訊,助於得到正確的結論。本文以臺灣扁柏苗木施肥試驗為例,以五個等級的氮肥為處理; 每隔四十天觀測苗高變化,重複觀測五次。以混合模式及一般線性模式做重複觀測的分析,將 時間效應分離,使處理效應顯現,分析結果顯示施肥效果有效,且時間與處理無交感,表示各 時段處理效應一致;若以單次苗高調查資料或淨生長量資料做變方分析及共變數分析,則不易 解釋不同時間內處理效應不一致的原因。不正確的統計模式可能帶來統計上不正確的推估,因 而導致錯誤的結論。 |
英文摘要 | Repeated measure design, especially longitudinal study, is often used, though not necessarily realized, in silvicultural research to understand treatment effects. When the design of an experiment is a repeated measure design, and the effect of time is not accounted for explicitly, the treatment and time effects may be confounded, and the treatment effect may therefore be erroneously estimated. In this study, using nursery fertilization of Taiwan cypress seedlings as an example, we analyzed the data from a repeated measure experiment using four different statistical models and compared their results. Analyses based on mixed model or general linear model, which accounted for the time factor explicitly in the models, suggested that the effects of fertilization were significant and consistent through time. However, when treating each repeated measurement as an independent trial, results from both regular analysis of variance and covariance models suggested that the effects of fertilization were not consistent through time, a consequence of specifying wrong models. |
本系統中英文摘要資訊取自各篇刊載內容。