数值变量资料的统计分析之统计描述专家讲座.pptx

学习目标和要求第一节数值变量资料统计描述组段划记频数f(1)(2)(3)二、集中趋势指标（掌握）（一）算数均数2.应用条件适合用于呈对称分布或近似对称分布资料。3.计算方法⑵加权法:n≥100例142.3156.6142.7145.7138.2141.6142.5130.5132.1135.5134.5148.8134.4148.8137.9151.3140.8149.8143.6149.0145.2141.8146.8135.1150.3133.1142.7143.9142.4139.6151.1144.0145.4146.2143.3156.3141.9140.7145.9144.4141.2141.5148.8140.1150.6139.5146.4143.8150.0142.1143.5139.2144.7139.3141.9147.8140.5138.9148.9142.4134.7147.3138.1140.2137.4145.1145.8147.9146.7143.4150.8144.5137.1147.1142.9134.9143.6142.3143.3140.2125.9①132.7152.9147.9141.8141.4140.9141.4146.7138.7160.9②154.2137.9139.9149.7147.5136.9148.1144.0137.4134.7138.5138.9137.7138.5139.6143.5142.9146.5145.4129.4142.5141.2148.9154.0147.7152.3146.6139.2139.9组段划记频数f(1)(2)(3)（二）几何均数2.适用条件：3.计算例⑵加权法:变量值个数较多或变量值为频数表资料例2-4（三）中位数和百分位数2.适用条件3.计算方法当变量值个数为偶数时计算公式为例例计算公式例潜伏期（小时）频数f累计频数累计频率（%）(1)(2)(3)（4）可确定中位数所在组段是12～，故L＝12，i＝6，fm＝38，n＝145，＝63。代入公式，得M=12+（6/38）×（145×50%－63）=13.5（小时）百分位数2．计算方法潜伏期（小时）频数f累计频数累计频率（%）(1)(2)(3)（4）练习题三、离散趋势指标（掌握）离散趋势指标（掌握）科学出版社卫生职业教育出版分社R甲=110－90=20（cm）R乙=104－96=8（cm）R丙=104－96=8（cm）缺点：①只用到最大、最小值，样本信息没能充分利用，不能反应全部数据变异程度。②受样本含量影响较大，样本例数越多，R可能越大，2组观察值例数悬殊时不用R比较。③样本含量不变时，每次抽样得到极差值相差较大，R稳定性较差。练习题四分位数间距方差样本方差用表示测定了5名健康人第一小时末红细胞沉降率，分别是6、3、2、9、10mm，求标准差。例142.3156.6142.7145.7138.2141.6142.5130.5132.1135.5134.5148.8134.4148.8137.9151.3140.8149.8143.6149.0145.2141.8146.8135.1150.3133.1142.7143.9142.4139.6151.1144.0145.4146.2143.3156.3141.9140.7145.9144.4141.2141.5148.8140.1150.6139.5146.4143.8150.0142.1143.5139.2144.7139.3141.9147.8140.5138.9148.9142.4134.7147.3138.1140.2137.4145.1145.8147.9146.7143.4150.8144.5137.1147.1142.9134.9143.6142.3143.3140.2125.9①132.7152.9147.9141.8141.4140.9141.4146.7138.7160.9②154.2137.9139.9149.7147.5136.9148.1144.0137.4134.7138.5138.9137.7138.5139.6143.5142.9146.5145.4129.4142.5141.2148.9154.0147.7152.3146.6139.2139.9练习题均数和标准差关系是A．均数愈大，s愈大B．均数愈大，s愈小C．s愈大，均数对各变量值代表性愈好D．s愈小，均数对各变量值代表性愈好答案：D变异系数例例练习题正态分布正态分布曲线正态分布特征（4）正态曲线下面积分布有一定规律。标准正态分布练习题医学参考值范围包括到采取单侧界值还是双侧界值问题，这通常依据医学专业知识而定。即不低于3.3L。2、百分位数法例2-1