复杂抽样调查设计二值资料一水平多重Logistic回归分析

本文目的是介绍复杂抽样调查设计二值资料多重logistic回归分析方法。通过一个实例,利用八种不同的分析策略(不考虑抽样设计和抽样权重、考虑抽样设计不考虑抽样权重、不考虑抽样设计考虑抽样权重、同时考虑抽样设计和抽样权重以及分别不考虑与考虑派生变量)对数据进行建模。对所得结果进行比较得出如下结论:在对复杂抽样设计资料进行统计分析的过程中,同时考虑抽样设计和抽样权重可以得到符合数据内部变量间依赖关系真实情况的结论。此外,本研究还介绍了采用SAS软件中SURVEYLOGISTIC过程对复杂抽样调查数据进行多重Llogistic回归分析的详细步骤。     

复杂抽样调查设计多值名义资料一水平多重Logistic回归分析

本文目的是介绍复杂抽样调查设计多值名义资料一水平多重logistic回归模型构建，并探讨不同策略之间的差异。采用SAS中的LOGISTIC过程和SURVEYLOGISTIC过程，分别按照是否考虑抽样设计与是否考虑抽样权重共4种分析策略对数据构建广义logistic回归模型，并比较结果。不同分析策略所得结果显示，不仅参数估计值、回归系数标准误、OR值及其置信区间的估计值有所差别，而且对纳入模型的解释变量也有影响。因此，在对复杂抽样调查设计多值名义资料构建广义logistics回归模型时，既要考虑抽样设计，又要兼顾抽样权重，否则即使样本量足够大，也会导致错误的推断结论。     

非配对设计多值名义资料一水平多重Logistic回归分析

本文目的是介绍非配对设计多值名义资料一水平多重logistic回归分析的基本原理、建模策略及注意事项。结合实例,应用SAS 9.4构建未经变量筛选和经变量筛选的多值名义资料多重logistic回归模型。通过回归分析的计算结果可知,同一变量的回归系数在不同logit函数中存在代数关系。多值名义多重logistic回归分析可以用来处理结果变量为多值名义变量的回归建模问题,并可以结合SAS实现对自变量的筛选,以获得简洁的回归模型。     

非配对设计多值名义资料多水平多重Logistic回归模型

本文目的是介绍非配对设计多值名义资料多水平多重logistic回归模型的构建与求解方法。首先介绍了有关的基本概念,涉及“多值名义结果变量”“分层或多水平数据结构”和“扩展的多重logistic回归模型”;其次,呈现了一个具有二水平结构的横断面调查资料,该资料涉及多个影响因素和一个多值有序的结果变量（在本文中,将其视为多值名义结果变量）;最后,借助SAS中的两个过程（即GLIMMIX和NLMIXED）对给定的资料进行统计分析,即构建和求解“非配对设计多值名义资料多水平多重logistic回归模型”,并对相关结果进行比较和解释。     

复杂调查资料的特点与统计分析方法概述

复杂抽样是在抽样过程中采用除一阶段单纯随机抽样外,其他抽样方法或其组合的抽样方案。本文对复杂抽样资料的特点、基于复杂调查资料进行差异性分析、多重回归分析以及进行生存资料多重回归分析的要点进行宏观概述。为科研工作者进行复杂抽样资料的分析提供参考和借鉴。     

非配对设计多值有序资料一水平多重Logistic回归分析

本文目的是介绍非配对设计多值有序资料一水平多重logistic回归模型的构建与求解方法。本文详细介绍了构建累积logistic回归模型的原理和具体方法,并结合实例介绍如何使用SAS软件中的LOGISTIC过程来拟合此回归模型,并对逐步回归法的输出结果进行了解释;其次讨论了有关构建累积logistic回归模型的过程中自变量筛选、模型评价以及拟合模型时需注意的问题。     

非配对设计多值有序资料多水平多重Logistic回归分析

本文目的是介绍多值有序资料多水平多重logistic回归分析方法。此法是在层次结构数据的基础上，构建多值有序因变量随一组自变量变化而变化的回归模型。具体的做法如下：①先介绍有关的基本概念；②呈现待分析的数据结构；③扼要介绍回归模型的构建与求解；④详细介绍如何使用SAS的GLIMMIX和NLMIXED两个过程来拟合此回归模型，并对相关结果进行解释和比较；⑤讨论多水平结构数据下拟合累积logistic回归模型时需注意的问题。     

非配对设计二值资料一水平多重Logistic回归分析

本文的目的是介绍非配对设计二值资料一水平多重Logistic回归模型的构建与求解方法。基于SAS软件分别对以列联表和数据库形式呈现的定性资料进行全面分析,并得出了4个对提高模型拟合优度很有价值的结论:第一,若资料以列联表形式呈现,应拟合"加权"Logistic回归模型;第二,若资料中包含定量自变量,不适合将其定性化;第三,若资料中包含定量自变量,应依据定量自变量和二值自变量产生出派生自变量;第四,若资料中有定性自变量时,必须将多值名义或有序自变量进行哑变量变换,不需要依据二值自变量产生出派生自变量。     

非配对设计二值资料多水平多重Logistic回归分析

本文目的是介绍非配对设计二值资料多水平多重logistic回归模型的构建与求解方法。首先介绍模型的有关概念及模型的构建原理,基于实例使用SAS软件对列联表资料进行分析,以proc glimmix和proc nlmixed过程构建和求解模型,并对相关结果进行解释和比较。     

如何正确运用χ~2检验——高维表资料统计分析方法概述

本文目的是介绍高维表资料的种类及其对应的统计分析方法。基于结果变量的资料类型,常见的高维表资料可分为以下三类,即结果变量为二值变量、多值名义变量和多值有序变量。高维表资料的统计分析方法主要有两大类,第一类为广义差异性分析,内容包括"加权χ~2检验""CMH χ~2检验"和"Meta分析";第二类为回归分析,内容包括"对数线性回归模型分析""Logistic回归模型分析""Probit回归模型分析"和"离散选择模型分析"。     

The association between pathological gambling and attempted suicide: findings from a national survey in Canada.

OBJECTIVE: To examine the association between pathological gambling (PG) and attempted suicide in a nationally representative sample of Canadians. METHODS: Data came from the Canadian Community Health Survey, Cycle 1.2, conducted in 2002, in which 36 984 subjects, aged 15 years or older, were interviewed. Logistic regression was performed with attempted suicide (in the past year) as the dependent variable. The independent variables were PG, major depression, alcohol dependence, drug dependence, and mental health care (in the past year), as well as a range of sociodemographic variables. Survey weights and bootstrap methods were used to account for the complex survey design. RESULTS: In the final logistic regression model, which included terms for PG, major depression, alcohol dependence, and mental health care, as well as age, sex, education, and income, the odds ratio for PG and attempted suicide was 3.43 (95% confidence interval, 1.37 to 8.60). CONCLUSIONS: PG (in the past year) and attempted suicide (in the past year) are associated in a nationally representative sample of Canadians. However, it is not possible to say from these data whether this represents a causal relation.     

Socioeconomic and Geographic Inequalities of Internet Addiction in Korean Adolescents

The aim of this study was to evaluate the social economic and geographic influences on Internet addiction in Korean youth using the Korean Youth Risk Behavior Web-Based Survey. Middle and high school students (n=73,238) were randomly selected from the respondents to a web-based survey using two-stage stratified complex sampling. A Geographical Information System was used to generate a Korean map of the level of Internet addiction for each province and complex sampling lower case logistic regression was used to investigate the relationship between social economic status (SES) and Internet addiction. Gyeonggi-do and Gyeongsang-namdo provinces have a higher total scores of Internet addiction. There were associations between Internet addiction and low SES (OR=1.504, 95% CI 1.156-1.956, p<0.01). There may need to be at multilevel approach which recognizes the different levels of need in provinces as well as trying to understand why the differences arise at an individual level.     

Are behavioral problems in childhood and adolescence associated with bipolar disorder in early adulthood?

Several recent studies have found an association between conduct problems and bipolar disorder in adolescents. However, prospective studies are rare and most do not apply multivariable analysis strategies to control for important variables (e.g. socio-demographics). The aim of this study was to test the association between certain conduct problems and bipolar disorders. The sample consisted of 591 adolescents (male and female) representative for 2,600 persons from the Canton of Zurich in Switzerland. Data were prospectively collected through an interviewing procedure, with the first screening taking place at the age of 19–20. The incidence rate was computed using sampling weights, and risk factors of bipolar II disorder were estimated using a multivariable logistic regression model. The 9-year incidence rate of bipolar II disorder in the canton of Zurich was 8.4% (n = 65). Adolescents and children showing behavior such as repeated running away from home and physical fighting were 2.6–3.5 times more likely to experience a bipolar II disorder than those with no indication of conduct problems. Sensitivity analysis showed that the conduct problems were not the result of low socio-economic status.     

Children with mental disorders in primary care: functional status and risk factors

Using data derived from a large epidemiological survey of children attending a pediatric primary care clinic, this article reports the relationship between the presence of mental disorders and functional status and the risk factors for depressive, anxiety-related and conduct disorders in these children. The results suggest that children with mental disorders in primary care are more functionally impaired than those without such disorders, with depressive and conduct disorders more often associated with such impairment. Analysis using logistic regression analysis to determine risk factors showed that broadly different familial and antecedent features predicted the presence of these groups of disorders. The findings confirmed the salience of factors relating to family and socioeconomic status and of maternal psychological health in a culture characterized by the extended family structure and widespread economic deprivation.