量表的选项顺序效应及其影响因素分析——以教育领域的李克特量表为例

An Analysis of the Order Effect of Options and Its Influencing Factors in the Scale:Taking Likert Scale in Education as an Example

李克特量表的选项顺序效应在教育领域是否存在具有争议,如果存在选项顺序效应,意味着研究者可以通过操纵选项影响研究结果。通过对878名教师作答的4个量表的A、B卷分析,发现选项顺序效应在教育领域量表中存在:1)选项5(非常符合)~1(非常不符合)排列的量表得分高于1(非常不符合)~5(非常符合)排列的量表得分;2)被调查者受首因效应的影响,更倾向选择第一个出现的选项,但选项1(非常不符合)~5(非常符合)排列的量表中,被调查者更容易选择否定含义的选项;3)在量表层面,所测概念指向他人、第三人称表述的题目越多、在问卷中位置越靠前以及反向表述的题目越少的量表越容易出现选项顺序效应;4)在具体题目层面,正向表述及在问卷中位置越靠前的题目,越容易出现选项顺序效应。选项顺序效应,具有明显中国人的认知特点,即受到整体型认知模式和高语境社会知识的影响,这值得研究者在设计问卷和量表时关注。

It is controversial whether the response order effect of Likert scale exists in the education field.If it exists,researchers can influence the research results by manipulating the responses.In this study,878 teachers finished a long questionnaire with four scales.The questionnaire is designed as format A whose responses displayed from 1(strongly disagree)to 5(strongly agree)and B whose responses displayed from 5(strongly agree)to 1(strongly disagree)to prove the existence of response order effect in education field.The scores were higher on format B than that on format A.Teachers were more likely to choose the first response,because of the primacy effect.However,in format A,teachers were more likely to choose the response with negative meaning.At the scale level,the scales in which the measured concepts referred to“others”,the scales which included more questions being expressed in the third person,the scales whose position were higher in the whole questionnaire,and the scales in which fewer questions were expressed negatively were more likely to appear the response order effect.In terms of specific questions,the questions expressed positively and the questions whose positions were higher in the whole questionnaire were more likely to appear the response order effect.The response order effect reflects obvious cognitive characteristics of easterners.It was influenced by holistic pattern of easterners and high-context culture.Therefore,researchers should pay attention to response order effect when designing questionnaires and scales.