资源描述
,Spearman 与Pearson,汇报人：屠星月,内容,Spearman相关系数的含义 Spearman相关分析的特点 Spearman 与Pearson相关分析的对比 结果的解释 对我的研究数据的适用性,一,Spearman相关系数的含义,Spearman rank-order correlation coefficient,两个连续或等级/秩变量之间的单调相关（ monotone association）关系。 单调函数描述两个变量之间的关系的程度（ how well the relationship between two variables can be described using a monotonic function） Pearson 相关系数的一个特例（系数计算中的X,Y被其秩代替）：,Rx and Ry are the ranks of the x and y variables,respectively,is the square of the difference between the corresponding ranks of xi and yi, and n is the number observations., 2,二,Spearman相关分析的特点,不对变量的分布做假设：非参数（自由分布） 可用于等级数据（ ordinal data：The most fundamental requisite is to be able to measure our observed correspondence by a plain numerical symbol ） 当数据分布导致，pearson相关系数不适用或者产生误导的使用。when the distribution of data makes Pearsons correlation coefficient undesirable or misleading 当数据为定性的变量，或者含有异常值时尤其适用。,三,Spearman 与Pearson相关分析的对比,三,Spearman 与Pearson相关分析的对比,三,Spearman 与Pearson相关分析的对比,1. 对同样的数据，Spearman相关统计相关，而pearson可能显著也可能不显著 the significance of Spearmans correlation can lead to the significance or non-significance of Pearsons correlation coefficient even for big sets of data 2.对同样的数据，二者所得的相关系数符号有可能不一样 （It is possible to meet a situation where Pearsons coefficient is negative while Spearmans coefficient is positive.） 3.注意：确保不要针对两个变量的相关强度，对spearman相关系数过度解释 Make sure not to overinterpret Spearmans rank correlation coefficient as a significant measure of the strength of the associations between two variables.,四,结果的解释,四,结果的解释,r 与P值,相关系数r（值和符号） 1- 置信区间 P值,呈现分析结果时应当包含的信息,当统计显著，但相关系数不大，则有可能是其他变量的影响,四,结果的解释,r 的大小,Benedict K，The Correlation Coefficient - Explained in Three Steps，,四,结果的解释,不要在非统计意义的“显著性”“显著地”这样的词 而用“统计上显著的”, 不要试图在P值的量级上进行判断。（to make clear that you do not intend to make any value judgement on magnitude when you use the term “significant”） When you refer to biological significance (when using effect sizes) use the expression biological relevance 统计显著不代表相关性强 （Keep in mind that a statistically significant result does not entail that the magnitude of the effect is relevant biologically (and that on top of that it is supported by statistics) “统计上显著的”仅仅代表P值小于预先设定的阈值 当P值高于 时，需保证已经提前做过power test，再接受原假设 r 不一定要很大才表示相关，应当基于知识判断生物学相关 （ Biological relevance is decided a priori and it is not necessarily a large value. You have to judge the magnitude you consider biologically relevant on a case by case basis.） 不确定性。对所有的估计参数都提供置信区间（这也是对P值的补充） 记住P值不是拒绝原假设的直接测量。P值越小，不代表“越好”。Only using Bayesian statistics you obtain a true measure of evidence.） P值只是表明，如果原假设为真时，获得现有数据的可能性。Remember that all your p-value provides you with is the probability of having obtained your data, or more extreme data, and only if the null hypothesis that you stated is absolutely true. No other interpretation of what a p-value means is correct.,六,参考文献,Sullivan, G. M. and R. Feinn (2012). “Using effect size-or why the P value is not enough.“ Journal of graduate medical education 4(3): 279-282. Support, M. E. “A comparison of the Pearson and Spearman correlation methods.“ Hauke, J. and T. Kossowski (2011). “Comparison of values of Pearsons and Spearmans correlation coefficients on the same sets of data.“ Quaestiones geographicae 30(2): 87-93. Martinez-Abrain, A. (2008). “Statistical significance and biological relevance: A call for a more cautious interpretation of results in ecology.“ acta oecologica 34(1): 9-11. Power of a Statistical Test - MoreS Benedict K，The Correlation Coefficient - Explained in Three Steps， Amanda Rockinson-Szapkiw，How to Calculate Statistical Power Using SPSS，,谢谢！,
展开阅读全文