[en] The Whisper-1.1 statistical analysis code targeted for nuclear criticality safety validation studies was released as part of the MCNP6.2 code release in 2018. Of the three primary functions that Whisper performs to calculating an application baseline upper subcritical limit, this work primarily studies the details of the benchmark selection process. From a catalogue of ICSBEP experiments included with Whisper, it is necessary to select which are most similar to an application of interest. Presently, the most common sensitivity/ uncertainty-based metric to help analysts in assessing similarity between an application and a benchmark model is the correlation coefficient, c_k. This quantity provides a metric for how the nuclear data induced uncertainty between the models are shared and alike. Because the c_k between an application and benchmark is reliant and sensitive to the nuclear data covariance information, further investigation into the properties and behavior of c_k is the topic of this paper. A fine-grained view into the c_k metric, how it is computed, which quantities most contribute to it, and potential alternative metrics to be aware of are discussed in detail. (authors)