Replication Data for: Is satisficing responsible for response order effects in rating scale questions?

Rating scales are used extensively in surveys. A rating scale can descend from the highest to the lowest point or from the positive to the negative pole. A rating scale can also start with the lowest point (or the negative pole) and ascend to the highest point (or the positive pole). Previous research has shown that the direction of the scale, i.e., the order of the response options, has an impact on responses, and that respondents are more likely to select response options close to the starting point of the scale, regardless of whether the scale starts with the lowest or the highest point. This paper advances the literature by examining empirically whether or not the response order effect in rating scale questions is driven by satisficing. Drawing on data from an experiment on five multi-item grids, we found that scale direction had a significant and extreme impact on response distributions. Although the effect of scale direction was stronger among speeders than non-speeders, the effect was observed across the board among those who were at a high risk of satisficing and those who were not.

评定量表（Rating scale）在调查研究中应用极为广泛。评定量表可设置为从最高分值向最低分值递减的形式，或是从正向端点延伸至负向端点的排布方式；也可采用以最低分值（或负向端点）为起点，向最高分值（或正向端点）递增的结构。既往研究表明，量表的排布方向——即应答选项的排序逻辑——会对受访者的作答结果产生影响：无论量表以最低分值还是最高分值作为起始点，受访者都更倾向于选择靠近量表起点的应答选项。本研究通过实证检验，探讨评定量表类问题中的应答顺序效应是否由满足式应答（satisficing）所驱动，以此拓展了该领域的研究文献。本研究依托一项包含五道多条目矩阵题的实验数据开展分析，结果发现量表排布方向对作答分布具有显著且极强的影响效应。尽管量表方向的影响效应在快速作答者（speeders）群体中较非快速作答者群体更为突出，但无论受访者是否存在高满足式应答风险，该效应均普遍存在于所有受试群体中。