“Throw it in as a Covariate?” Common Problems Using Measured Control Variables in Experimental Research
讨论了实验会计和审计研究中协变量使用的常见陷阱，帮助作者、审稿人和编辑避免或纠正这些错误，提高研究可信度。
SUMMARY This paper discusses common pitfalls in the use of covariates in experimental accounting and auditing research. Sometimes these mistakes are uncovered in the review process, giving authors more problems to wade through (including a little credibility loss with reviewers) in an already-difficult review process. Sometimes these mistakes are not uncovered in the review process, leading to analyses that support dubious inferences in the published literature. Sometimes these mistakes are even introduced into papers by reviewer comments. Two objectives of this paper are (1) to help authors and coauthors (including early-career researchers) prevent some common, unforced errors from making their way into their research, and (2) to give authors, reviewers, and editors a reference for explaining within the review process why a covariate should be modeled in a certain way, without having to begin the discussion of these ideas from scratch.