Scholarlink

<p>Previous research highlights the impact of nonverbal cues and visual layouts on respondents’ interpretations and answers to survey questions (Schwarz et al. in Pub Opin Q 55(4):570–582, 1991; Tourangeau et al. in Pub Opin Q 71(1):91–112, 2007; Toepoel and Dillman in Soc Sci Comput Rev 29(2):193–207, 2011). In this study, we investigate the effect of numerical values on responses to five questions measured using 11-point bipolar scales. These questions asked for evaluations of the general impact of climate change and its effects on different groups. Respondents were randomly assigned to scales ranging from 0 to 10, or from − 5 to + 5. The scales were presented horizontally and had identical verbal endpoints anchored as “extremely negative” and “extremely positive”. The experiment was embedded in a cross-national face-to-face survey conducted in Hungary, Portugal, and Great Britain. Contrary to prior research (Schwarz et al. in Pub Opin Q 55(4):570–582, 1991; O’Muircheartaigh et al. J Off Stat 11:295–308, 1995; Tourangeau et al. in Pub Opin Q 71(1):91–112, 2007), our analyses show an opposite trend: the means tend to be consistently and significantly lower for the − 5/+5 scale. The discrepancy between our findings and previous studies suggests that the interpretations of the 0/10 and − 5/+5 scales could be item specific. The implications of these findings and how they fit with previous work are discussed.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Between two minds: the influence of numerical labels on survey responses in a cross-national study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月31日

查看全文
  • Gianmaria Bottoni,
  • Eva Aizpurua

摘要

Previous research highlights the impact of nonverbal cues and visual layouts on respondents’ interpretations and answers to survey questions (Schwarz et al. in Pub Opin Q 55(4):570–582, 1991; Tourangeau et al. in Pub Opin Q 71(1):91–112, 2007; Toepoel and Dillman in Soc Sci Comput Rev 29(2):193–207, 2011). In this study, we investigate the effect of numerical values on responses to five questions measured using 11-point bipolar scales. These questions asked for evaluations of the general impact of climate change and its effects on different groups. Respondents were randomly assigned to scales ranging from 0 to 10, or from − 5 to + 5. The scales were presented horizontally and had identical verbal endpoints anchored as “extremely negative” and “extremely positive”. The experiment was embedded in a cross-national face-to-face survey conducted in Hungary, Portugal, and Great Britain. Contrary to prior research (Schwarz et al. in Pub Opin Q 55(4):570–582, 1991; O’Muircheartaigh et al. J Off Stat 11:295–308, 1995; Tourangeau et al. in Pub Opin Q 71(1):91–112, 2007), our analyses show an opposite trend: the means tend to be consistently and significantly lower for the − 5/+5 scale. The discrepancy between our findings and previous studies suggests that the interpretations of the 0/10 and − 5/+5 scales could be item specific. The implications of these findings and how they fit with previous work are discussed.