I've spent the week after vacation catching up on my journal referee chores (two down, four to go [but only two are late, I think]). At the risk of revealing myself as your anonymous referee**, here is a comment that I seem to be increasingly making when I review stated preference papers:
It is best practice to present a data summary for the dependent and independent variables. This is important for several reasons (in no particular order). First, the data summary is needed to determine the efficacy of the modelling approach. Second, patterns in the univariate data across different stated preference scenarios should be described before the empirical model is presented so that readers can better visualize what is being modeled and better assess what the model is doing. Third, means of independent variables may be used along with the empirical model in benefit transfer of elasticities and consumer surplus. Fourth, the magnitude of the variation in the dependent variable is needed to understand the effect sizes measured by the regression coefficients. Fifth, given the large benefits and the low cost of a data summary table, failure to present one raises questions about the validity of the data.
Have I missed any reason to report your data summary? And what are the reasons why you'd want to not report it (other than intentionally hiding something)?
Why does it seem more and more acceptable to present your empirical model in Table 1? Given Retraction Watch and increasing scrutiny about data quality and replication, I would think that one would want to be as forthcoming as possible about the data that you are working with. Indeed, you have gone to considerable effort and expense to collect stated preference data. Go ahead and show it off in a data summary table.
*You thought I was going to say beer post, didn't you?
**If you have gotten this paragraph in the past or get it in the future, remember that haters gonna hate.
