<ul data-eligibleForWebStory="false"><li>Data scientists often face challenges in translating fuzzy concepts into concrete target variables for predictive modeling tasks.</li><li>A study involving interviews with fifteen data scientists in education and healthcare reveals that they construct target variables through a bricolage process.</li><li>The target variable construction process involves iterative negotiation between high-level measurement objectives and practical constraints to satisfy criteria like validity, simplicity, predictability, portability, and resource requirements.</li><li>Data scientists employ adaptive problem (re)formulation strategies, such as swapping target variables or combining multiple outcomes, to meet modeling objectives effectively.</li></ul>

Measurement as Bricolage: Examining How Data Scientists Construct Target Variables for Predictive Modeling Tasks

Discover more