“My paper provides scientific tests of CMV, and finds evidence that refutes it, leading to my conclusion that CMV is nothing more than an urban legend.”

In psychology and related fields, people assume that the use of a single method automatically inflates observed relationships among measured variables. We see this issue raised most often with survey studies using questionnaires, where people report about themselves and their experiences, because it is assumed the use of the method itself caused variables to be related. This is called the common method variance (CMV) or mono-method bias problem.

My paper provides scientific tests of CMV, and finds evidence that refutes it, leading to my conclusion that CMV is nothing more than an urban legend. It is cited by researchers in responding to reviewers and editors who raise the CMV concern. It also is cited by others who have written on this and related methodological issues.

It describes tests for something assumed to be true, and finds disconfirming evidence. I argue that the impact of measurement bias is far more complex than the simple-minded idea of CMV. Each measured variable can be affected by a unique set of factors depending on both the method used and the nature of the construct of interest. Some factors will cause biases, that is, the measure will be affected by something not intended.

For example, people can exhibit a self-serving bias and over-rate themselves if asked how well they perform. Such biases can be shared among two or more variables, but not among all variables. Although we can be secure that CMV is not a problem that we need to worry about, we do need to be concerned about shared biases that are more difficult to detect and deal with than CMV would have been if it existed.

My paper shows that something believed by scientists to be true, is not.

Researchers frequently have their papers rejected because of the CMV criticism. A little over 20 years ago, a doctoral student who had this happen asked me for advice in dealing with this criticism of his work. It made me wonder what evidence might exist for CMV, which started me doing research to find it. I have yet to see credible evidence that it exists, and I have seen evidence which suggests it does not.

I have a broader interest in how we can draw conclusions from data in nonexperimental field studies. I can see myself doing more work on strategies that can be used to deal with this problem.