This article specifcally notes the researchers' awareness of "the concern for the economic health of bars" and in their statements they repeatedly stress they examained both restaurants AND bars.

The authors of this study had two sepaarate sets of NAICS employment data:  Data for bars, AND data for full-service-restaurants.  They deliberately chose NOT to publish any analysis on what happened to bar employment.  

Some might conclude that they simply withheld such results because it would have damaged the political cause the researchers and their antismoking funders, ClearWay Minnesota, intended to support.   There’s an obvious motivation to have performed such a separate analysis since the results, if they went the "right" way, would have made the study’s conclusion FAR more powerful! After all, just picture the headlines: "New Study Shows Even BARS Gain Business After Smoking Bans!"

When I asked the study’s lead author about the lack of separation I was simply told that the analysis of both together was "the most appropriate" approach.  I then noted that I found this puzzling unless the researchers were fairly new to the field since historically it’s always been thought that bars suffer disproportionately.  I asked, “Are you saying such a separation and its value to your study did not occur to you and your colleagues and that none of you or your peer-reviewers / journal editors thought to take a look at that data?"

Her response to me, instead of offering a reasonable explanation, was this:

"You may want to familiarize yourself with some of the scientific literature on economic effects on the hospitality industry."

She then attached an old study blaming any contrary research results on tobacco industry funding and corruption.  This seemed a rather odd response since my own research has always been quite explicitly unfunded.

This is not what I would call a professional response to a reasonable question. Picture if I did a study on the economic meltdown, examining the data for its effects on Blacks and Whites. I know the Black community has a better reason for concern and has also been most featured in the media as suffering, but despite having both sets of data I simply decide it is "most appropriate" to combine the numbers for Blacks and Whites and present a report concluding that there was no harm from the meltdown.  I do this despite knowing that, since Whites outnumber Blacks ten to one, any Black suffering will of course be covered up.  

I then present my study to the media, arguing there's no need for government change by saying, "We certainly did not detect anything close to the dramatic claims that opponents make based on the concerns that they have for Blacks." (That’s an actual newspaper quote from the lead researcher with the word Blacks substituted for Bars.) The headlines the next day then read: “New Research: Economic Meltdown Does Not Hurt Blacks, Whites."

And when asked why I didn’t examine Black suffering separately I simply reply that I felt combining the data was “more appropriate” while telling my questioner to study economic history and handing them a rather questionable racial pamphlet.

My opinion about researchers who would do such a thing may seem harsh, but I fail to see much, if any, path between the choices of incompetent or unethical. The fact that my emails did not elicit a response of "Gee, we ARE new to the field and just didn't think of separating bars." would seem to strongly imply the latter point of view.

The fact that the study cited only two pieces of contrary research, with one being an old radio broadcast and the other one being cited improperly (the only improper citing among all 36 references), lends support to that view.

The fact that the research compared the three main variables in an extremely odd way that lumped two of them together quite awkwardly gives more support to that view.

The fact that this study ignored an earlier study by the same organization that showed drastic customer reductions in 7 out of 10 bars in publicly shared data, refused to release the rest of their data, and seem to have somehow even removed the shared data from regular Internet channels adds even further support.

And the final fact, the fact that reasonable questions were met with noncommunicative responses, seems to complete the perception of ethical questionability in the case of this study.

Any legislature considering this research in their decision-making process regarding a smoking ban needs to go back and take a very hard look both at it, and at all the other questionable data they have been given.  I have done extensive research and writing in this field, and I can assure you: This study does not stand alone in its practices.


Michael J. McFadden
Author of "Dissecting Antismokers' Brains"