Measuring Attitude with Positively Packed Self-Report Ratings: Comparison of Agreement and Frequency Scales

Abstract

Self-report rating scales with balanced response formats, anchored with vague frequency of activity indicators, often elicit inadequate information, especially when respondents are inclined toward a generally positive attitude toward the psychological object being rated. This study used an unbalanced or positively packed rating scale with both frequency and agreement response anchors within the context of a questionnaire about studying and learning practices and conceptions for high school students ( N=734). Psychometric characteristics and communication factors were investigated using 12 pairs of items for which both frequency and agreement response formats were used. Communication factors identified by Schwarz in 1999 such as small changes in wording provided adequate explanation for changes in response rate or fit to the IRT measurement model for three pairs of items. Although psychometric evidence of the superiority of agreement over frequency response format was not conclusive, continued use of agreement anchors with a positively packed rating scale appears justified.