Me parece que el potencial psicométrico de un único ítem es nulo. Los mas recientes estudios psicométricos en personalidad y psicopatía coinciden en modelos complejos, multicategóricos y continuos. Esto último se está llevando a la medición de síntomas en enfermedades mentales para, posiblemente, poder identificar subtipos fenomenológicos en algunas de ellas.
There were a few outstanding papers in the 1980's on psychometric models for single items. This paper provides details of modelling strategies for single items
JAGODZINSKI, W., & KÜHNEL, S. M. (1987). Estimation of Reliability and Stability in Single-Indicator Multiple-Wave Models. Sociological Methods & Research, 15(3), 219–258. https://doi.org/10.1177/0049124187015003003
..and was followed by another that showed how longitudinal models can be developed for single items.
Article Is there a "Socratic Effect" in Nonexperimental Panel Studie...
Indeed,
Indeed, Willem Saris has written extensively on this issue for many decades and his work has been very influential. Here's a good starting point-
Quality and Quantity
November 1990, Volume 24, Issue 4, pp 471–490
Common factors can always be found but can they also be rejected?