Common method variance in NEO - FFI and IPIP personality measurement

Including a method factor on which all items loaded significantly improved fit of CFAs of both IPIP and NEO-FFI data. Fit was further improved by separating the method factor into two factors representing construct-positive and construct-negative items. The criterion-related validity of conscientiousness increased when uncontaminated by method bias. PRESS PARAGRAPH This research extends previous findings of contamination of personality scores by common method bias to the popular NEO-FFI measure of personality traits. Previous studies have found that responses to the IPIP Five Factor Model personality assessment are contaminated by common method bias, and that factors representing constructpositive and construct-negative items contribute separately to that contamination. The present research found similar results for the NEO-FFI measure, indicating almost 14 percent of total variance in responses is associated with such response biases. The effects of this contamination on validity are shown and suggestions for dealing with it are discussed.