Tests for Outliers

This training program will address when to exclude outlying data points from an analysis. It will illustrate how best to apply sensitivity analysis to determine the impact of removing data. Participants will learn to perform and interpret hypothesis tests for outlier detection.

Why Should You Attend:

When analyzing data, the decision whether to exclude specific data points as non-representative outliers is an important one. The exclusion of one or more flyers may have a very large impact on the results of the analysis. While it is tempting to simply remove points that appear to be different, doing so may cover-up real issues that must be addressed.

This webinar will provide guidelines for determining whether removing suspect data points is justified or not. If the decision is not clear, the use of sensitivity analysis to determine the effect of the outlier(s) on the results of the analysis will be described. Additionally, several statistical outlier tests will be covered to provide objectivity to the decisions. Numerous examples will be provided to illustrate the decision making process and various outlier tests.

Learning Objectives:

• Understand when to exclude outlying data points from your analysis
• Apply sensitivity analysis to determine the impact of removing data
• Perform and interpret hypothesis tests for outlier detection

Areas Covered in the Webinar:

• Defining outliers
• Probability distributions
• Determining whether or not to exclude specific data points
• Sensitivity analysis
• Statistical tests for outliers (e.g. Grubbs, Dixon)
• Interpreting hypothesis tests for outliers
• Documenting data removal

Who Will Benefit:

• Business Analysts
• Product/Process/Business Managers
• Quality and Process Engineers
• Plant Operations Personnel
• Engineers
• Scientists
• R&D Personnel
• Product Development personnel
• Personnel utilizing data to make decisions