This study meta-analyzed IRT item discrimination parameter estimates and CFA item loadings to explore the methodological effect of negatively... Show moreThis study meta-analyzed IRT item discrimination parameter estimates and CFA item loadings to explore the methodological effect of negatively-worded items in personality measures. We found three important moderators that determined whether the negatively-worded item effect affected a scale. The first moderator was the manner in which the scale was defined. We found a strong negatively-worded item effect for Neuroticism but not for Emotional Stability. The personality scale was also a moderator, with a negatively-worded item effect being observed for Agreeableness, Extraversion, Neuroticism, and Openness but a positively-worded item effect for Emotional Stability and Conscientiousness. Third, low-motivation samples tended to produce a larger negatively-worded item effect. Moreover, there was no statistically significant difference between personality and non-personality inventories regarding the negatively-worded item effect. Finally, item negation did not produce the expected effect. Practical implications and limitations of the study are discussed. Ph.D. in Psychology, May 2015 Show less
The study was designed to investigate how item writing would affect psychometric properties of personality measurements. New personality items... Show moreThe study was designed to investigate how item writing would affect psychometric properties of personality measurements. New personality items were developed based on ideal point and dominance models, which were analyzed in the generalized graded unfolding model (GGUM), and two personality scales were constructed. Writing dominance items were slightly easier (more successful) than writing ideal point items, but this varied slightly by personality dimensions. Of three tactics for writing ideal point items trying to identify neutral items was less effective than appealing to an average or using double-barreled items. Scales constructed using ideal point items had substantially inferior psychometric properties as compared to scales created using dominance items, including lower reliability, lower validity and mixed test information results. However, lower validity of ideal point scales may be due to lower reliability of the measurements. Practical and methodological implications were also discussed in the paper. M.S. in PSYCHOLOGY, December 2012 Show less
