当前位置: X-MOL 学术Research in Social and Administrative Pharmacy › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Towards a reconsideration of the use of agree-disagree questions in measuring subjective evaluations
Research in Social and Administrative Pharmacy ( IF 3.7 ) Pub Date : 2021-06-24 , DOI: 10.1016/j.sapharm.2021.06.014
Jennifer Dykema 1 , Nora Cate Schaeffer 1 , Dana Garbarski 2 , Nadia Assad 3 , Steven Blixt 4
Affiliation  

Agree-disagree (AD) or Likert questions (e.g., “I am extremely satisfied: strongly agree … strongly disagree”) are among the most frequently used response formats to measure attitudes and opinions in the social and medical sciences. This review and research synthesis focuses on the measurement properties and potential limitations of AD questions. The research leads us to advocate for an alternative questioning strategy in which items are written to directly ask about their underlying response dimensions using response categories tailored to match the response dimension, which we refer to as item-specific (IS) (e.g., “How satisfied are you: not at all … extremely”). In this review we: 1) synthesize past research comparing data quality for AD and IS questions; 2) present conceptual models of and review research supporting respondents' cognitive processing of AD and IS questions; and 3) provide an overview of question characteristics that frequently differ between AD and IS questions and may affect respondents’ cognitive processing and data quality. Although experimental studies directly comparing AD and IS questions yield some mixed results, more studies find IS questions are associated with desirable data quality outcomes (e.g., validity and reliability) and AD questions are associated with undesirable outcomes (e.g., acquiescence, response effects, etc.). Based on available research, models of cognitive processing, and a review of question characteristics, we recommended IS questions over AD questions for most purposes. For researchers considering the use of previously administered AD questions and instruments, issues surrounding the challenges of translating questions from AD to IS response formats are discussed.



中文翻译:


重新考虑在衡量主观评价中使用同意/不同意问题



同意-不同意(AD)或李克特问题(例如,“我非常满意:强烈同意……强烈不同意”)是衡量社会科学和医学科学中态度和意见最常用的回答格式之一。本综述和研究综述重点关注 AD 问题的测量特性和潜在局限性。这项研究促使我们提倡另一种提问策略,其中编写项目来直接询问其潜在的响应维度,使用适合响应维度的响应类别,我们将其称为特定于项目(IS)(例如,“如何您满意吗:一点也不……非常满意”)。在这篇综述中,我们:1)综合了过去的研究,比较了 AD 和 IS 问题的数据质量; 2) 提出支持受访者对 AD 和 IS 问题认知处理的概念模型并回顾研究; 3) 概述 AD 和 IS 问题之间经常存在差异并可能影响受访者认知处理和数据质量的问题特征。尽管直接比较 AD 和 IS 问题的实验研究产生了一些混合结果,但更多研究发现 IS 问题与理想的数据质量结果(例如有效性和可靠性)相关,而 AD 问题与不良结果相关(例如默许、响应效应等) .)。根据现有的研究、认​​知处理模型以及对问题特征的回顾,我们在大多数情况下建议使用 IS 问题而不是 AD 问题。对于考虑使用先前管理的 AD 问题和工具的研究人员,讨论了围绕将问题从 AD 转换为 IS 回答格式的挑战的问题。

更新日期:2021-06-24
down
wechat
bug