Explore chapters and articles related to this topic
System Stability and Sustainability
Published in R. S. Bridger, Introduction to Human Factors and Ergonomics, 2017
Common method variance or common method bias occurs when the same instrument is used to measure different outcomes. For example, suppose we ask 100 people to rate on a 5-point scale whether they experience role conflict at work and how dissatisfied they are with their jobs. We find that the scores on the two measures correlate highly—those with high role conflict are also highly dissatisfied. According to measurement theory, the variance in the scores consists of true variance and random error variance; for example, VARjob dissat=VARtrue+VARrandom
The Validity of Self-reported Traffic Behaviour Data
Published in Anders AF WÅhlberg, Driver Behaviour and Accident Research Methodology, 2017
The first problem is simply how well any given method actually measures what it is intended to measure. For self-reports, this could be the reported number of traffic accidents as a reflection of the real number. Common method variance is a term applied to associations within data sets that are not due to any real correspondence, but to a common way of measuring different variables. Self-reports are notorious for such effects, although this has hardly been noted within traffic safety research. One example of a simple mechanism that can affect self-reports is a scale response bias, where some respondents tend to use the extremes of the scales, whereas others tend to reply in the middle range, despite their actual behaviours/attitudes/experiences being the same. Within a sample, this will create an artefactual association between variables. This kind of self-report validity threat will be described further in a section designated to its effects within traffic safety research.
Role of Internet Self-Efficacy and Interactions on Blended Learning Effectiveness
Published in Journal of Computer Information Systems, 2022
Ritanjali Panigrahi, Praveen Ranjan Srivastava, Prabin Kumar Panigrahi, Yogesh K Dwivedi
Common Method Bias (CMB) is caused due to the instrument of data collection rather than the appropriate representation of the construct items. Since the data is collected through survey methodology from the respondent one at a time, it is likely to introduce common method bias.91 The presence of common method bias can inflate or deflate the correlations among variables in the data. Hence, it is crucial to take procedural measures before data collection to minimize the possible CMB in actual data collection.92 A psychological separation is created by asking response variable immediately and giving time to answer the rest of the variable at a different time. Further, to minimize the common method bias, the respondents’ anonymity is protected which reduced the evaluation apprehension. Moreover, the scale items are improved by defining unfamiliar or ambiguous terms and keeping the items simple and concise. Additionally, a statistical test for common method variance is conducted to check whether the majority of variance is explained by a single factor. Harman’s single-factor test is conducted in SPSS to detect the presence of common method variance as it is accepted as good statistical criteria for CMB.93 The result showed that no single factor accounted for more than 50% (36.846%) of the variance. Thus, no general factor is identified that explains the majority of the variance in data. This implies that common method variance is not likely to influence the results of the study.
Adoption of green supply chain management practices in multi-tier supply chains: examining the differences between higher and lower tier firms
Published in International Journal of Production Research, 2022
Seongtae Kim, Kai Foerstl, Christoph G. Schmidt, Stephan M. Wagner
We sought to ex-ante reduce the impact of common method variance on the results of our study through the careful design of the survey tool. As recommended by Podsakoff et al. (2003), we first queried the control variables, followed by the independent and mediating variables before asking respondents to report on the dependent variable. Moreover, common method bias was assessed by applying Harman’s single-factor test. Estimating a confirmatory factor analysis (CFA) revealed that the single-factor model did not fit our data well (χ2=762.47; χ2/df=8.47; CFI=0.718; RMSEA=0.162). The CFA including all our latent constructs (theory-based) provides significantly better model fit indices (χ2=132.147; χ2/df=1.65; CFI=0.978; RMSEA=0.048). Additionally, we found that modelling the marker variable loadings (Marker: We must frequently change our marketing practices to keep up with the market and competitors) onto the substantive (theory-based) indicators does not improve the model fit significantly, indicating no common method bias (theory-based vs. marker-based model) (Δχ2=.02; Δdf=1; χ2critical=2.97; no difference) (Williams, Hartman, and Cavazotte 2010). Overall, we conclude that in this study, common method bias does not significantly affect our results.
How social media use is related to student engagement and creativity: investigating through the lens of intrinsic motivation
Published in Behaviour & Information Technology, 2022
Muhammad Awais Gulzar, Mudaser Ahmad, Marria Hassan, Muhammad Imran Rasheed
Our findings should be seen in the light of its limitations. For example, our study used the cross-sectional research design, which has a threat of common method variance (CMV). We have taken procedural remedies such as keeping the confidentiality and anonymity of respondents and statistically testing to ensure that there is no such issue of CMV. However still, we believe that a longitudinal research design or an experimental laboratory study may enhance the validity of testing our model. Second, our model tested intrinsic motivation as a mediating variable between the association of social media utilisation by the students and its optimistic outcomes. Future scholars can find alternative explanations for this relationship. For example, students’ self-efficacy may be one reason why students’ utilisation of social media is allied with positive results in academia. Third, future researchers can explore other ultimate outcomes of social media usage by students, such as well-being and career exploration. Fourth, we tested cyberbullying as the boundary condition in our model; future researchers can identify some other moderating variables in this model. A moderating variable at the second stage of our model can be of great value. Finally, as we conducted our study in a single country (a collectivistic culture), testing of our model in other contexts, countries, and cultures can generate different findings. Future researchers can, therefore, test our model in other countries and regions of the world with different cultures. Particularly testing in an individualistic culture will be of great interest.