MinitabUG 发表于 2022-7-25 11:06:47

假设检验的功效和样本数量

在假设检验中,我们会使用样本中的数据来描绘有关总体的结论。首先,我们会进行假设,这被称为原假设(以 H0 表示)。当您进行原假设时,您也需要定义备选假设 (Ha),其与原假设正相反。样本数据将用于判断 H0 是否可以被否定。如果其被否定,则统计结论将认为备选假设 Ha 正确。 请记住这一检验的功效,或是在原假设不正确时,原假设被否定的可能性。 它可以解释为“检验在应该拒绝原假设时拒绝原假设的能力”。如果原假设不正确,则有很高概率拒绝原假设是很有意义的。功效与类型 2 的错误相关(功效 = 1 - 类型 2 错误),请见下表。类型 2 错误是当备选假设正确时不拒绝原假设的概率。因此,确保有足够高的功效,才能保证类型 2 错误较低或“可以接受”。确保检验有足够功效的一种常用方法是收集足够的数据,因为功效的计算取决于样本数量等因子。样本数量越大,功效越高。换言之,未能收集足够的数据将导致低功效和大量类型 2 错误。 data:image/png;base64,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 最重要的是要找到合适的样本数量。显而易见,未能收集足够的数据会导致更多的类型 2 错误。但是,收集“过多”的数据也会增加类型 1 错误,因为检验的功效会更高。因此,该检验可能会检测到与假设值的微小差异,即使该差异可能没有任何实际意义,尤其是与抽样成本有关时。检验功效的计算应当基于实际意义。
MINITAB具有通过多种不同统计检验计算功效的功能 在下列示例中,分析人员在 Minitab 中通过单比率检验和单样本 t 检验,进行了功效和样本数量分析。 单比率检验样本数量
考虑将产品分类为好或差的制造过程,其中有 1% 的不良品率。如果不良品率上升至 3%,则会对整个组织造成严重的成本问题。他们需要确定合适的样本数量,以满足:类型 I 错误率为 0.05,检验功效为 0.80,以检测出不良品率从 1% 上升至 3% 或更高。 因为分析人员对不良品率研究感兴趣,他们使用了单比率检验。原假设和备选假设是: Ho: P = 0.01 Ha: P > 0.01 其中 P 为实际缺陷比率。为了找出需要多少数据点才能达到至少0.8的功效,分析人员在 Minitab 中进行了单比率检验的功效和样本数量分析。 data:image/png;base64,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 单样本 t 检验的样本数量 将产品分类为好或差很简单,但会损失很多信息。将好产品视为在 5 到 10 之间。假如有 2 个单元测得的数值为 4.9 和 10.01,并因而归入差的分类。假如有另外 2 个单元测得的数值为 2.3 和 14.1,并因而归入差的分类。请注意,如果只是简单的区分好和差,这两种情况是相同的。因此,如果测量产品质量特征是可行的,那么分析人员应该记录质量特征的实际值,并使用记录的数据 – 无需转换为好和差。单样本 t 检验可以用于检验总体的均值是否与目标一致。如果样本数据的均值接近“目标”,则该过程可能运行良好。如果均值不接近目标,则可能生产出缺陷产品。 例如,假设该产品特征为特定目标的孔直径。分析人员可以测量每个产品上的孔直径,并使用单样本 t 检验将均值与目标值进行比较,而不是检查 236 个产品以确定孔是否符合规格。
为了找出需要多少数据点来检测至少 80% 功效的过程均值中的 1 西格玛偏移,分析人员在 Minitab 中对一个单样本 t 检验进行功效和样本数量分析。 data:image/png;base64,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计算的样本数量仅为 10。这意味着如果分析人员希望确定目标的偏离均值是否超过了 1 西格玛,则他们需要进行 10 个单位的单样本 t 检验,以获得至少 80% 的功效。 为什么会有这么大的区别? 属性数据的假设检验需要大量样本,因为在收集数据时没有获取详细信息。另一方面,连续数据的假设检验只需较少的样本数量,因为其获取并使用了产品的详细信息。该理论不仅适用于功效。属性数据需要大量样本以用于置信区间、属性一致性分析、控制图和能力分析。 总之,重要的是进行具有足够功效的假设检验,以提供合理的机会来检测差异。功效与样本数量直接相关。Minitab 具有计算多种不同假设检验(包括试验设计)的功效的功能。 本文最初出现在Minitab博客上。
页: [1]
查看完整版本: 假设检验的功效和样本数量