Automatic Implicit Motive Codings Are at Least as Accurate as Humans’ and 99% Faster

Abstract

Implicit motives, nonconscious needs that influence individuals’ behaviors and shape their emotions, have been part of personality research for nearly a century but differ from personality traits. The implicit motive assessment is very resource-intensive, involving expert coding of individuals’ written stories about ambiguous pictures, and has hampered implicit motive research. Using large language models and machine learning techniques, we aimed to create high-quality implicit motive models that are easy for researchers to use. We trained models to code the need for power, achievement, and affiliation (N = 85,028 sentences). The person-level assessments converged strongly with the holdout data, intraclass correlation coefficient, ICC(1,1) =.85,.87, and.89 for achievement, power, and affiliation, respectively. We demonstrated causal validity by reproducing two classical experimental studies that aroused implicit motives. We let three coders recode sentences where our models and the original coders strongly disagreed. We found that the new coders agreed with our models in 85% of the cases (p