摘要
使用已经发表的4组生境选择数据计算Bonferroni和Bailey’s置信区间,以比较分析Bailey's方法在研究"利用-可利用型"生境选择数据时的优势。用Bailey’s方法对4组数据中的两组做出了与Bonferroni方法不同的统计推断,特别是当Bonferroni方法对其中的一组数据无法对动物是否具有生境选择性做出判断时,而Bailey’s方法却发现了差异。如同Cherry(1996)指出的,Bonferroni方法是基于拟合优度卡方检验的,要求大样本的独立数据,当出现小样本时会因为连续性和一致性缺失而做出错误推断。Bailey’s方法因为使用了连续性校正因子克服了这一缺陷,而且使用时无须进行拟合优度卡方检验,是一种简便可靠的生境选择数据分析方法。建议在推荐Bailey’s方法的同时,应适当增大样本量、控制同时分析的生境类型数量,以控制分析时I类和II类的错误率。
This paper calculated the Bonferroni interval and Bailey＇s interval of four sets published data in order to show comparative dominance of Bailey＇s method in the study of ＂use-available＂ model on habitat selection of animals. Bailey＇s method deduced differently on two sets aforesaid data compared to Bonferroni method, one dataset of which couldn＇t show whether the animals made habitat selection or not in Bonferroni method＇s judgment. Nevertheless Bailey＇s method detected the variance, Cherry （1996） pointed out that the Bonferroni method of interval construction is based on a large sample approximation for a confidence interval for a single binomial proportion with the overall confidence level controlled with the Bonferroni inequality. This method will lead to a logically wrong deduction because of lacking coherence and consonance for small samples. As an alternative, Bailey＇s method calculates correctly with the use of a continuous correcting factor to overcome the above-mentioned limitation, furthermore a Chi-square goodness-of-fit test is undesired. We recommended Bailey＇s method in analyzing data of habitat selection. We also suggested that researchers should properly increase the sample size meanwhile decrease the number of type of habitats to control Type Ⅰ and Type Ⅱ errors.
基金
国家自然科学基金(30470235)
内蒙古自治区高等学校科学研究项目(NJzy08164)
