Outliers in Response Data

An outlier is a numerical response to a TSC KPI that is much larger or smaller than expected. The outlier may exist because the responding supplier has a significantly better or worse process than their peers, leading to extra-ordinary outcomes, or the outlier response may represent a measurement error.

In all three cases, it’s useful to identify those responses which can be labeled as outliers. TSC uses both empirical methods (IQR, z-score) and data gathered based on empirical research from peer-reviewed journal articles (i.e., published GHG emissions intensities) to identify outlier responses. Characteristics of the dataset being evaluated determine which method is used to identify outliers.

This process occurred at the end of the THESIS campaign with suppliers being notified if TSC identified any potential outliers in their assessments and have three weeks to respond.