LEAN SIX SIGMA WORLD CONFERENCE
Overcoming Bias – A Core Skill for Effective Data Collection
Presenter: Russ R. Aikman, LSS Program Manager, TMAC/The University of Texas at Arlington, Arlington, Texas, USA
Keywords: Bias, Data, Collection, Representative, Sample, Error
Fundamental to Lean Six Sigma is data-based decision making. Before data can be used for problem-solving, LSS practitioners must ensure it is truly representative of the process from which it was obtained. All Black Belts and Green Belts learn critical concepts for data collection including operational definitions, sampling techniques (randomization, stratification, systematic), and measurement system analysis. Less well understood is the concept of bias, and the negative impact it can have on sound data collection. What exactly is bias, and how can it be overcome? This presentation will include a definition of bias including how it may be calculated. Some of the most common forms of bias will be discussed such as self selection, self exclusion, judgement, convenience, grouping, and missing key representatives. Less well known types of bias will also be shared including survivor bias. Finally, some general guidelines on overcoming bias in data collection will be shared.