Book

Common method variance or measurement bias? The problem and possible solutions

Publisher: Sage
19 Followers
 · 
137 Reads
    • "Different sources of method effects are suggested, including the form or the length of the response scales and the form or complexity of the questions. We agree that CMV and CMB can be found in survey data, especially in the case of self-report surveys, because we understand that the respondent as a rater is a source of method variance; that is, variance not intended by the researcher (Spector & Brannick, 2009). However, we do not believe that this CMV stems from the method itself given that the latter is a constant. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Since the idea of method variance was inspired by D. T. Campbell and Fiske in 1959, many papers have demonstrated an ongoing debate about both its nature and impact. Often, method variance entails an upward bias in correlations among observed variables—common method bias. This article reports a split-ballot multitrait–multimethod experimental design for estimating 2 opposite biases: the upward biasing method variance from the reaction to the length of the response scale and the position of the survey items in the questionnaire and the downward biasing effect of poor data quality. The data are derived from self-reported behavior related to emotional and social competencies. This article illustrates a methodology to estimate common method bias and its components: common method scale variance, common method occasion variance, and the attenuation effect due to measurement errors. The results show that common method variance has a much smaller impact than random and systematic measurement errors. The results also corroborate previous findings: the greater reliability of longer scales and the lower reliability of items placed toward the end of the survey.
    No preview · Article · Aug 2014 · Structural Equation Modeling A Multidisciplinary Journal
    • "We do not resolve this dispute in our research, because other papers handle this discussion very well. See, for example, Spector and Brannick (2009) or Podsakoff et al. (2003). Overall, our research is well aligned with the perspective that a method is a mechanism where by an item triggers unrelated personality traits within the respondent. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This research analyzes the effects of common method variance (CMV) on parameter estimates in bivariate linear, multivariate linear, quadratic, and interaction regression models. The authors demonstrate that CMV can either inflate or deflate bivariate linear relationships, depending on the degree of symmetry with which CMV affects the observed measures. With respect to multivariate linear relationships, they show that common method bias generally decreases when additional independent variables suffering from CMV are included in a regression equation. Finally, they demonstrate that quadratic and interaction effects cannot be artifacts of CMV. On the contrary, both quadratic and interaction terms can be severely deflated through CMV, making them more difficult to detect through statistical means.
    No preview · Article · Jul 2010 · Organizational Research Methods
    • "There have been many studies since D. T.Campbell and Fiske (1959)on method effects where multitrait– multimethod (MTMM) is the most commonly used design for estimating method effects. Lance, Dawson, Birkelbach, and Hoffman (2010) provided an overview of these MTMM studies done in psychology.Saris and Gallhofer (2007)in the case of self-report surveys, because we understand that the respondent as a rater is a source of method variance; that is, variance not intended by the researcher (Spector & Brannick, 2009). However, we do not believe that this CMV stems from the method itself given that the latter is a constant. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Personality and behavioural assessment are common practice in research and applications throughout the world. Most of this research has been in English-speaking countries. Relatively little work has been done in cross-cultural settings to study the nature of response scale. Familiarity with an 11-point response scale in European countries may affect the reliability of questionnaire responses. In Spain, with mainly European MBA students, results show that an 11-point scale provides composites with greater reliability, validity and less invalidity, than the commonly used 5-point scale.
    No preview · Article · Apr 2009 · Personality and Individual Differences
Show more