There are several approaches for treating with common method bias including pre-data-collection strategies as well as post-collection statistical analysis. In the Ringle program SmartPLS, what is the appropriate procedure that is to be employed?
Dear Silburn,
unfortunately, most post-hoc approaches to deal with common method bias have been turned out to be not very effective, see
Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010). On making causal claims: A review and recommendations. The Leadership Quarterly, 21, 1086-1120.
Conway, J. M., & Lance, C. E. (2010). What reviewers should expect from authors regarding common method bias in organizational research. Journal of Business Psychology, DOI 10.1007/s10869-010-9181-6.
Richardson, H. A., Simmering, M. J., & Sturman, M. C. (2009). A tale of three perspectives: Examining post hoc statistical techniques for detection and correction of common method variance. Organizational Research Methods, 12(4), 762-800.
As far as I know there are two options to address common method bias:
a) Multi-trait-multi-method approaches
b) instrumental variable methods where the instrument(s) can be plausibly regarded as not targeted by the common method.
By the way: Its unfortunate to call this thing common method bias: First: a "method" is no cause (rather the response style to the method), and second: it should be called common source/cause bias as it is a typical case of a common cause bias that results from both the IV and DV are incluenced by a common cause - may it be an ommited variable, response style to the method or any other common cause.
Best,
Holger
I think smartpls can only use pre-data-collection strategies.
The updated practice of assessing CMB may refer to them:
Chin, W.W., Thathcer, J.B., and Wright, R.T. 2012a. “Assessing Common Method Bias: Problem with the ULMC Technique," MIS Quarterly (36:3), pp. 1003-1019.
Chin, W.W., Thathcer, J.B., Wright, R.T., Steel, D.J. 2012b. “Controlling for Common Method Variance in PLS Analysis: The Measured Latent Marker Variable Approach,” Proceedings: 7th International Conference on Partial Least Squares and Related Methods, Houston, Texas USA, pp. 1-8.
Dear Silburn,
unfortunately, most post-hoc approaches to deal with common method bias have been turned out to be not very effective, see
Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010). On making causal claims: A review and recommendations. The Leadership Quarterly, 21, 1086-1120.
Conway, J. M., & Lance, C. E. (2010). What reviewers should expect from authors regarding common method bias in organizational research. Journal of Business Psychology, DOI 10.1007/s10869-010-9181-6.
Richardson, H. A., Simmering, M. J., & Sturman, M. C. (2009). A tale of three perspectives: Examining post hoc statistical techniques for detection and correction of common method variance. Organizational Research Methods, 12(4), 762-800.
As far as I know there are two options to address common method bias:
a) Multi-trait-multi-method approaches
b) instrumental variable methods where the instrument(s) can be plausibly regarded as not targeted by the common method.
By the way: Its unfortunate to call this thing common method bias: First: a "method" is no cause (rather the response style to the method), and second: it should be called common source/cause bias as it is a typical case of a common cause bias that results from both the IV and DV are incluenced by a common cause - may it be an ommited variable, response style to the method or any other common cause.
Best,
Holger
Hi Mark and Holger,
Appreciate the comments and the references provided. Have done a review and the references have been very helpful. I have also since broadened my investigation to look at covariance-based SEM solutions and not just PLS. Specifically the addition of a Common Latent Factor (CLF) latent variable in AMOS looks promising.
Silburn
Hi Silburn,
Yup, there are more methods to handle CM bias in AMOS.
Mark
Dear Silburn,
do you mean with a CLF a additional factor on which all indicators of all other latents load? If yes, I disagree as this method has beein shown (see my provided literature) not to solve the CMB problem. Would be nice....
You should incorporate all imaginable third variables which could act as common causes and control them. Further, better measures the IV and DV from different sources (if this is reasonable). And finally, you can use instrumental variables as predictors of your IV's (see the Antonakis paper). All these approaches rely on assumptions (namely that all relevant third variables are included) but there is no causal inference without assumptions.
The only thing we can do is to increase the evidence and not to proof the model. I don't think, however, that pure response styles have an enormous biasing effect, but I am open to be corrected. It's the substantial variables that have to be controlled....
Best,
Holger
Hi Holger,
Yes, I was indeed suggesting that by including a CLF parameter in the AMOS analysis I would be able to assess and mitigate CMB.
On the basis that CMB is driven from a bias due to same instrument, same population group, same time data collection, I will be adjusting my sampling strategy with time-separated collection of the variables as well as using an independent population group for the DV
Thanks
Silburn
Maybe you can conduct marker variable using common method bias. This test also useful
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling