Abstract
Described herein is the academic lineage and independent validation of the Self-Efficacy Teaching and Knowledge Instrument for Science Teachers-Revised (SETAKIST-R). Data from 334 K-12 science teachers were analyzed using Partial Credit Rasch models. Principal components analysis on the person-item residuals suggest two latent dimensions: Knowledge and Teaching Self-Efficacies. Item-fit statistics were used to select items for each subscale. Person and item separation (reliability) indices were quite low, and we noted disordered response patterns on the person-item maps that revealed problems with item content and/or scaling for both subscales. These issues include the presence of: verbal negatives, ambiguous modifiers, counter-intuitive scaling, and an "undecided/uncertain" option. The SETAKIST-R, in its current form, cannot be recommended as a measure of science teacher self-efficacy.