统计方法

分析与建模

Gaining knowledge and insights are the goals of 分析与建模. 趣赢平台 statisticians are well-versed in a variety of analytic and modeling techniques, ranging from simple linear regression to complex multivariate linear and log-linear models, 层次贝叶斯模型, 倾向和预测模型, 随意的推理, 多层次建模和小面积估计. These techniques are often used in conjunction with statistical data integration or simulation methods. Modeling techniques can be applied to a multitude of data types, 比如基于概率或非概率的样本, 行政记录, 观察性或基于设计的数据. 趣赢平台 works in collaboration with clients to support transparency and reproducibility of analytic results, providing documentation and annotated programs of the applied methodology.

数据协调和数据链接

Bringing together data from disparate sources can help answer complex questions or solve complicated problems. Data harmonization is frequently needed to unify disparate data fields, formats, or dimensions. 使用统计方法进行数据协调, 比如标准化和归算, 允许更大的数据效用. Data linkage methods are used to connect records or units across the different sources. 趣赢平台 brings a depth of knowledge and experience in these areas to support client 分析, 研究, 以及基于证据的决策.

数据隐私

趣赢平台 is a highly respected and trusted industry leader in statistical confidentiality and privacy protection. Our statisticians apply a variety of rigorous methods to quantify the risk of disclosure (using statistical models and record linkage), develop data treatments for risk mitigation (including consideration for a formal privacy framework), and to provide creative data access solutions (such as a flexible table generator). To ensure confidentiality and privacy protection of respondent data, 趣赢平台 works collaboratively with our clients to balance the risk of disclosure with the usefulness of the final data products.

调查统计

Since our inception in 1963, 趣赢平台 has been world-renowned for expertise in survey statistics. We bring innovative statistical approaches throughout the entire process, 从框架创建开始, 复杂的样品设计和样品选择, 数据收集和提取, 数据准备和清理, 归责, 权重, 无反应偏倚分析, 估计和方差估计, 数据的获取和传播. 趣赢平台 statisticians are adept at applying these methods to traditional populations of interest, 比如人, 家庭, 或企业, 以及新的种群, 如出版物的语料库, 电子健康记录（EHRs）, 或者路段. 目的是设计出高质量的产品, 具有成本效益的, 用于研究的代表性数据产品, 分析, 和决策.