单因素重复测量设计与随机区组设计资料处理异同解析

目的探讨单因素重复测量设计与随机区组设计资料的方差分析的关系。方法利用文献数据^[1,2]，拟对资料进行单因素重复测量设计和随机区组设计2种处理，比较2种处理结果的异同。结果单因素重复测量设计资料在满足球对称的条件和正态对称条件下，与误用随机区组设计处理结果相同；尤其在球不对称且F值＞＞F界值时，采用随机区组设计处理与单因素重复测量分析校正前后的结果，也会出现一致情况。结论2种方法都是单因素试验的方差分析，但2种设计的方法和分析目的均不同，不可混同。     

交叉试验设汁统计方法探讨

有些临床试验设计是在病人之间比较不同处理的效果，其常用的统计方法如团体，检验．完全随机化设计方差分析等，而另一些临床试验设计是在病人内部对不同处理进行比较，常用的统计分析方法如配对，检验，随机区组设计方差分析等。     

不均衡配对资料的统计分析

一、问题的提出要比较某种处理因素(或调查因素,下同)的两个水平下的某种变量值的差别,如比较病人和正常人的某种生理指标值或生化指标值的差别,可用两组资料的 t 检验或配对资料的 t 检验。但配对的 t 检验在每个配对区组内只能是1∶1配对,一般要求配对区组数较多,用实验或调查收集资料比较困难;而且不能研究处理因素和区组因素之间的交互作用。本文     

多元析因设计的方差分析及其SAS程序实现

[目的]针对医学研究中常见的多元析因设计方差分析存在的问题,给予统计软件技术上的支持。[方法]利用SAS统计软件包,结合实例,编写程序,计算相应的统计量。[结果]运用SAS程序对实例数据由单因素分析及多因素分析,得出实例数据中的两处理因素及其交互作用均有统计学意义。[结论]研究临床中多元析因设计利用SAS软件包进行统计分析的途径,为类似问题的解决提供了参考。     

裂-裂区设计及其方差分析

目的介绍裂-裂区设计及其方差分析。方法结合实例介绍裂-裂区设计中实验单位、实验因素级别的正确识别和合理安排,以便正确选择方差分析模型。结果只有正确识别裂-裂区设计中实验单位、实验因素的级别,才能获得正确的方差分析模型以及相应的分析结果。结论裂区及裂-裂区设计是采用区组化控制技术进行的多因素研究设计,是完全随机区组设计、拉丁方设计等的综合运用。     

缺失数据配伍组设计方差分析的回归分析方法

本文介绍一种缺失数据配伍组设计方差分析的精确方法——回归分析方法,该方法不同以往使用的估项方法,其实质是把方差分析模型视为一般线性模型,采用边界条件法处理,也适用于配伍组设计方差分析非缺项情形。我们将回归方法和估项方法同原来未缺项的方差分析结果对比,认为回归方法比估项方法准确,尤其在处理组结果上,回归方法比估项方法更准确。     

重复测量设计与随机区组设计原理及应用

重复测量设计与随机区组设计是医学研究领域两种重要的研究设计类型,二者在形式上常较为接近,尤其是当重复测量或区组因素均与时间效应有关时,更易使人产生混淆[1].在实际工作中,两者误用的情形并不少见.如此不仅会损失原始数据的部分信息(如重复测量数据内部的相关性),甚至有可能导致统计分析结论的错误.因此本文就重复测量设计与随机区组设计的自身特点及其区别作一番介绍,并运用实例对两种设计类型的SPSS软件分析结果进行分析和比较.     

Friedman M检验平均秩的多重比较在SPSS软件的实现

完全随机区组设计的秩和检验(Friedman's test)是随机化区组设计方差分析不满足方差分析条件时采用的方法.随机化区组设计的秩和检验是由M·Friedam在符号检验的基础上提出来的,又称M检验,目的是推断各处理组样本分别代表的总体分布是否不同[1].对于Friedman M检验,在当P＜α(α为检验水准)差异有统计学意义时,可认为多个总体间相比较有差异,但不能说明任何两个总体间均是有差异的.     

不等距重复测量设计方差分析用SAS和SPSS实现的对比

目的探讨不等距重复测量设计方差分析的条件及结果解释。方法用SAS和SPSS软件实现不等距重复测量设计方差分析。结果给出了方差分析条件、SAS和SPSS程序及主要结果解释。结论只有满足方差分析条件、正确合理解释结果,才能保证不等距重复测量设计方差分析的有效性。     

拉丁方实验的方差分析简化法

拉丁方是目前常用的一种实验设计方法,有一次性和重复性两种类型,其实验结果的统计学分析均采用方差分析法。鉴于拉丁方实验具有各因素(行间、列间、处理间、重复批间)划区内的例数相等的特点,可利用电子计算器将其方差分析方法简化。简化法的基本公式还适用于那些按因素划区时各区内例数相等的完全随机、区组随机、正交设计等实验资料的方差分析。一、基本公式设拉验丁方实验为k×k级,重复次数为m,则可     

各类医疗机构空气消毒效果评价

目的:评价辖区内各类医疗机构空气消毒效果,指导空气质量监控重点。方法:采用卡方分析和随机区组设计的方差分析方法,将各类医疗机构分为公立、企事业单位和个体3组,每组采集空气样本30宗,评价空气消毒各格率和细菌总数指标的意义。结果:公立、企事业单位和个体医疗机构空气合格率分别为50·000%、56·667%、30·000%,P>0·05,无统计学意义。从细菌总数方差分析显示组间的作用有统计学意义,P=0·017,认为在分组不同的条件下测得的细菌总数不同。多重比较结果显示,前二者空气细菌总数结果之间的差异无统计学意义,但前二者分别与后者之间的差异有统计学意义。结论:细菌总数指标能较好地反映空气消毒效果,个体医疗机构是空气质量监控重点。     

Optimizing occupational exposure measurement strategies when estimating the log-scale arithmetic mean value--an example from the reinforced plastics industry

When assessing occupational exposures, repeated measurements are in most cases required. Repeated measurements are more resource intensive than a single measurement, so careful planning of the measurement strategy is necessary to assure that resources are spent wisely. The optimal strategy depends on the objectives of the measurements. Here, two different models of random effects analysis of variance (ANOVA) are proposed for the optimization of measurement strategies by the minimization of the variance of the estimated log-transformed arithmetic mean value of a worker group, i.e. the strategies are optimized for precise estimation of that value. The first model is a one-way random effects ANOVA model. For that model it is shown that the best precision in the estimated mean value is always obtained by including as many workers as possible in the sample while restricting the number of replicates to two or at most three regardless of the size of the variance components. The second model introduces the 'shared temporal variation' which accounts for those random temporal fluctuations of the exposure that the workers have in common. It is shown for that model that the optimal sample allocation depends on the relative sizes of the between-worker component and the shared temporal component, so that if the between-worker component is larger than the shared temporal component more workers should be included in the sample and vice versa. The results are illustrated graphically with an example from the reinforced plastics industry. If there exists a shared temporal variation at a workplace, that variability needs to be accounted for in the sampling design and the more complex model is recommended.     

Chronic Toxicity of Atrazine to Sago Pondweed at a Range of Salinities: Implications for Criteria Development and Ecological Risk

The objective of this study was to conduct a series of 28-d partial life-cycle atrazine subchronic toxicity tests with sago pondweed, Potamogeton pectinatus, at salinities of 1, 6, and 12 ppt. These data will be used for development of a chronic estuarine criterion for atrazine in Maryland waters of Chesapeake Bay and to determine possible ecological risk for a sensitive nontarget species. The three endpoints used for this submerged aquatic macrophyte were final wet weight, final dry weight, and final number of rhizome tips at the termination of the 28-d test. Dry weight was determined to be the most sensitive endpoint. Chronic values from a one-way ANOVA using dry weight were 21.2, 21.2, and 10.6 μg/L at salinities of 1, 6, and 12 ppt, respectively. Chronic values using wet weight were 21.2 μg/L at all three salinities. A chronic value of 94.9 μg/L was reported at all three salinities using rhizome tips as the endpoint. A two-way ANOVA was also used for analysis of data to increase the power of detecting differences among treatments and assess salinity interaction. The salinity effects were averaged in the two-way ANOVA. Both rhizome tips and dry weight were used in the two-way ANOVA; wet weight did not satisfy the equal variance assumption. The chronic value for rhizome tips was 94.9 μg/L atrazine, which is the same value reported from the one-way ANOVA. There was no salinity effect and no interaction between salinity and atrazine concentration. The chronic value determined from the two-way ANOVA using dry weight was 5.3 μg/L; there was a salinity effect but no significant salinity/atrazine concentration interaction. Salinity was not reported to effect either dry weight or rhizome tips after 28-d exposures in the controls but wet weight was significantly lower at 12 ppt when compared with 1 and 6 ppt. Ecological risk to sago pondweed from atrazine exposure was judged to be low based on recent exposure data from the mainstem, tributaries, and streams in Chesapeake Bay. Received: 4 September 1996/Accepted: 9 March 1997     

Model mis-specification and overestimation of the intraclass correlation coefficient in cluster randomized trials

Intraclass correlation coefficient (ICC) estimates must be provided when reporting the results of a cluster randomized trial. This study demonstrates that estimating this parameter with one-way ANOVA and an underlying mixed-effects statistical model leads to biased estimates. The bias depends on the effect size of the studied treatment.     

定量数据考虑基线与否的几种方差分析模型的模拟比较

目的比较几种考虑基线与否的方差分析模型的统计性能。方法应用Monte Carlo技术,在基线均衡和不均衡情况下,比较以下方差分析模型:以基线为协变量的变化量协方差分析(ANCOVA)、变化率协方差分析(PCS-ANCOVA)和对数变化率协方差分析(logPCS-ANCOVA);不考虑基线的变化量方差分析(ANOVA)、变化率方差分析(PCS-ANOVA)和对数变化率方差分析(logPCS-ANOVA)。以I类错误与检验效能评价各种方法的统计性能。结果在基线均衡的情况下,PCS-ANCOVA和ANOVA均可很好地控制I类错误,且检验效能都较高;在基线不均衡的条件下,若基线对因变量无影响,ANCOVA与ANOVA均可以较好地控制I类错误,此时ANOVA的检验效能高于ANCOVA;若基线对因变量有影响时,只有ANCOVA可以很好地控制I类错误,且检验效能较高,其他方法效果不佳。结论考虑到实际应用中绝大部分情况是基线对因变量有影响,即相关,建议优先采用以基线为协变量的协方差分析或变化量的协方差分析,无论基线是否均衡。用变化率做方差分析或协方差分析,有可能冒着比值的分布不满足参数方法条件的风险,应用时应慎重。     

便携式动态心肺功能测定仪的信度评价

目的：对k4b^2便携式动态心肺功能测定仪的信度进行评价。方法：利用k4b^2便携式动态心肺功能测定仪，采用重测法，随机抽取大学生中年龄在20－25岁的15名健康男性，在统一测试方法和统一测试状态（静息状态）下，进行3次重复测试，对试验数据进行单因素方差分析、变异度分析及正态性检验，比较其重测信度。结果：对测试对象总体而言，各指标的3次重复测定结果无显著性差异；对每个测试对象而言，按15％标准，各指标的变异系数中仅有1例略超过15％，其余均在15％以内，表明符合重测信度的要求；由试验数据可以计算健康成年男子心肺功能指标的参考值范围，但由于样本例数太少，需进一步扩大具有代表性参考值。结论：k4b^2便携式动态心肺功能测定仪具有良好的重测信度。     

Comparison between telephone and self-administration of Short Form Health Survey Questionnaire (SF-36)

OBJECTIVE: The characteristics of the 36 item Medical Outcome Short Form Health Study Survey (SF-36) questionnaire, designed as a generic indicator of health status for the general population, allow it to be self-administered or used in personal or telephone interviews. The main objective of the study was to compare the telephone and self-administered modes of SF-36 for a population from Girona (Spain). METHODS: A randomized crossover administration of the questionnaire design was used in a cardiovascular risk factor survey. Of 385 people invited to participate in the survey, 351 agreed to do so and were randomly assigned to two orders of administration (i.e., telephone-self and self-telephone); 261 completed both questionnaires. Scores were compared between administration modes using a paired t test. Internal consistency and agreement between modalities were analyzed by respectively applying Chronbach's alpha and intraclass correlation coefficients. The effect of the order of administration on the test-retest difference was analyzed by one-way ANOVA for repeated measurements. RESULTS: Physical function, physical role and social functioning received significantly lower scores when the self-administered questionnaire was used prior to the telephone survey. When the initial survey was conducted by telephone, all Chronbach's alpha coefficients (except social functioning) scored over 0.70 in the self-administered modality. The intraclass correlation coefficient ranged from 0.41 to 0.83 for the telephone-self order and from 0.32 to 0.73 for the self-telephone order. No clinically significant effect was observed for the order of application. CONCLUSIONS: The results of the present study suggest that the telephone-administration mode of SF-36 is equivalent to and as valid as the self-administered mode.     

Effects of model misspecification in the estimation of variance components and intraclass correlation for paired data

Paired data occur in many experimental situations. When one views the subjects as a random sample from some large population, it may seem reasonable to model the data according to the typical one-way random effects analysis of variance (ANOVA). It is then usually of interest to estimate variance components and intraclass correlation. These estimators can be biased if key assumptions are violated, leading to erroneous interpretations and conclusions. We focus upon assumptions about the equality or inequality of means and/or variances of the two measures on each subject. In the framework of the one-way random effects ANOVA model, and three generalizations of it, we document estimators obtained as solutions to the likelihood equations. We consider the potentially serious effects of mistaken assumptions. Our findings suggest that the most general model considered is most desirable if consistent and efficient estimation of the between-subject variance component and intraclass correlation is the main goal. We also briefly connect our exposition to the study of reliability or agreement.