Test & Evaluation of Optic Systems Using Survival Analysis
Oct 10, 2018 10:44 AM
| Last Modified: Oct 31, 2018 6:21 AM
Kevin Eng, Douglas Ray
In this study, the performance of digital and direct-view optic systems were compared under a controlled experimental setting.
Performance was measured based on the length of time a participating soldier took to identify a target. The data generated from the experiment contained many censored events where study participants were unable to identify a target in the allotted time. Previous analysis of similar studies done by ATEC use basic linear regression and ANOVA. These tools are powerful in the right context, but do not properly account for censored data, and therefore cannot accurately characterize the effects of factors of interest.
This poster provides an overview of survival analysis techniques used to account for censored data, and make credible comparisons between the direct-view and digital optic systems.