Holm Søren, Hofmann Bjørn
1 University of Oslo, Norway.
2 The University of Manchester, UK.
J Empir Res Hum Res Ethics. 2017 Oct;12(4):199-205. doi: 10.1177/1556264617714658. Epub 2017 Jul 14.
A precondition for reducing scientific misconduct is evidence about scientists' attitudes. We need reliable survey instruments, and this study investigates the reliability of Kalichman's "Survey 2: research misconduct" questionnaire. The study is a post hoc analysis of data from three surveys among biomedical doctoral students in Scandinavia (2010-2015). We perform reliability analysis, and exploratory and confirmatory factor analysis using a split-sample design as a partial validation. The results indicate that a reliable 13-item scale can be formed (Cronbach's α = .705), and factor analysis indicates that there are four reliable subscales each tapping a different construct: (a) general attitude to misconduct (α = .768), (b) attitude to personal misconduct (α = .784), (c) attitude to whistleblowing (α = .841), and (d) attitude to blameworthiness/punishment (α = .877). A full validation of the questionnaire requires further research. We, nevertheless, hope that the results will facilitate the increased use of the questionnaire in research.
减少科研不端行为的一个前提条件是了解科学家的态度。我们需要可靠的调查工具，本研究旨在调查卡里奇曼的“调查问卷2：研究不端行为”的可靠性。该研究是对斯堪的纳维亚地区生物医学博士生的三项调查（2010 - 2015年）数据的事后分析。我们进行了可靠性分析，并采用拆分样本设计进行探索性和验证性因素分析作为部分验证。结果表明，可以形成一个由13个项目组成的可靠量表（克朗巴哈α系数 = 0.705），因素分析表明有四个可靠的子量表，每个子量表衡量一个不同的构念：（a）对不端行为的总体态度（α = 0.768），（b）对个人不端行为的态度（α = 0.784），（c）对举报行为的态度（α = 0.841），以及（d）对应受责备/惩罚的态度（α = 0.877）。对该问卷的全面验证需要进一步研究。不过，我们希望研究结果将有助于该问卷在研究中得到更多使用。
