差異分數使用之省思－信度、解釋量、構念意義與衡量等同性的探討

A Reflection on the U sage of Difference Scores: The Discussion on Reliability, Interpretation, Construct Meaning, and Measurement Equivalence

熊欣華(Hsin-Hua Hsiung)

《調查研究》 10期 (2001/10) Pp. 67-103

https://doi.org/10.7014/SR.2001100003

摘要

差異分數是廣泛應用於社會與行為科學領域中的一種研究方法，主要在於評估構念間的一致或不一致。過去學者對差異分數的使用有一些方法上的爭議，這些爭議主要可分為四方面：信度、解釋量、構念意義、以及衡量等同性。本文認為不同類型的差異分數，在這四方面所產生的問題嚴重性並不相同。當研究屬於「變化的衡量」時，方法上的主要問題僅在於信度，此時情況比較單純，但是當研究的衝量的對象與構念不同時，在這四方面都可能出現問題。差異分數雖然有方法上的問題，但在某些研究中有其必要性。對於不同方面的限制，本文提出可能的改善之道與評估方式。此外，提出幾種替代方法，分別為「肉體成長模型」、「區別項目函數」、「直接比較法」與「反應面模型法」，並說明其應用狀況。希望藉由這些討論，使研究者對差異分數有所思考，以便未來有較佳的應用。

關鍵字

差異分數 ； 信度 ； 衡量等同性 ； 內隱成長模型 ； 反應面模型法

並列摘要

Difference scores are widely used in social science researches and are aimed to evaluate the consistency between constructs. However, there are some methodological controversies regarding the usage of this research method. These controversies are categorized in four aspects: reliability, interpretation, construct meaning, and the measurement equivalence of the difference scores. This study attempts to show that the degrees of urgency of the problems encountered in these four aspects are not the same. It varies in different kinds of difference scores. For instance, if the study is focused on measuring change by time, then the researcher should focus on the aspect of reliability. Fortunately, this problem is not very serious. On the other hand, if the study is meant to measure different subjects and different constructs, then the researcher must deal with the problems in all four aspects.Although problems as to the usage of difference scores exist, it is still imperative for some studies. This paper tries to provide some ways to lessen the amount of difficulties involved and evaluates them. There do possess, however, other alternatives that researchers may avail of, such as ”latent growth modeling”, ”differential item functioning”, ”direct comparison”, and ”response surface model”. These methods have their unique application contexts. Thus, researchers may think twice on difference scores in order to apply them properly in studies.

並列關鍵字

difference scores ； reliability ； measurement equivalence ； latent growth modeling ； response surface model

參考文獻

Caplan, R. D.(1987).Person-Environment Fit Theory and Organizations: Commensurate Dimensions, Time Perspectives, and Mechanisms.Journal of Vocational Behavior.31,248-267.

Chan, D.,Schmitt, N.(2000).Interindividual Differences in Intraindividual Changes in Proactivity During Organizational Entry: A Latent Growth Modeling Approach to Understanding Newcomer Adaptation.Journal of Applied Psychology.85(2),190-210.

Chiou, J. S.,Spreng, R. A.(1996).The Reliability of Difference Scores: A Re-Examination.Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior.9,158-167.

Cronbach, L. J.(1958).Proposals Leading to Analytic Treatment of Social Perception Scores.Person Perception and Interpersonal Behavior.353-379.

Cronbach, L. J.,Furby, L.(1970).How We Should Measure 'Change'-or Should We?.Psychological Bulletin.74(July),68-80.