Detecting Effects of Positively and Negatively Worded Items on a Self-Concept Scale for Third and Sixth Grade Elementary Students

Method effects associated with item wording have been explored in a variety of instruments and found that the practice of using positively- and negatively- worded items may introduce systematic measurement errors that disrupt analyses and interpretations of the results. Therefore, the first purpose in the present study was to explore if method effects were present in a Chinese general self-concept scale, originally developed in Chinese. The second purpose was to determine if the factor structure of the method effects, if present, differed for third and sixth grade students. The third purpose was to determine if the observed method effects were related to other substantively meaningful variables. Results from a series of CFAs support the presence of method effects associated with the negatively and positively worded items and method effects were largest for the negatively worded items. The results from multigroup model comparisons indicate that the factorial structure of these method effects was not significantly different for third and sixth graders. Three demographic variables, including student gender, student grade level, and students' overall performance ratings provided by teachers, were used to examine the relationships with negative method effects. The results of path analysis indicated that students who were rated lower by their teachers were more likely to endorse negative statements about themselves. Furthermore, students in grade 3 were significantly more likely to endorse negative statements compared to students in grade 6. But gender was not significantly related to the negative method factor. (Contains 8 figures and 6 tables.)