Volume: 17  Number: 1  Page: 45–60

Profiles of Respondents Who Respond Inconsistently to Positively- and Negatively-worded Items on Rating Scales

Gail H. Weems1, Anthony J. Onwuegbuzie2 and Daniel Lustig3 1University of Arkansas Little Rock, USA, 2University of South Florida, USA and 3University of Memphis, USA

Many instruments, especially Likert-type scales, contain both positively- and negatively-worded items within the same scale (i.e. mixed-item format). A major reason for this practice appears to be to discourage response sets from emerging. Using this format also helps the analyst to detect response sets that occur in data sets, and thus eliminate them from subsequent analyses. However, some psychometricians seriously question the use of mixed-item formats, positing that positively- and negatively-worded items within a scale are not measuring the same underlying trait. Limited evidence has suggested that certain individuals are more predisposed to providing differential response patterns when responding to a mixed-item format scale. However, to date, only a few characteristics of these differential-responding individuals have been identified. Thus, the purpose of this present study was to extend this line of research. Specifically, the researchers analysed responses to several scales utilizing mixed-item formats. For example, on canonical correlation analysis, a sample of 185 students revealed a relationship between degree of differential responses between positively- and negatively-worded items on three 6-item measures of foreign language anxiety (i.e. input, processing, and output anxiety) and several dimensions of self-perception, study skills, and locus of control. Implications of all findings are discussed.

Keywords: rating scales, participant characteristics, reverse coding, itemwording surveys

© 2003 G.H. Weems et al.

Full text PDF