I am inclined to agree with Dr. Meyer here.
I would recommend reserving ASA objections and keeping our partisan powder dry to address very egregious behavior that fundamentally impacts the ability of scientists to do their work and the basic reliability of scientific data, behavior whose nature more resembles the Andreas Georgeiou prosecution.
While an ongoing study has separate issues, once a study and its report are final, when to publicize it is a management or political matter rather than a scientific one, and any mistiming issue do not impact the results’ fundamental scientific reliability or integrity. I think this means it’s not our issue.
Statisticians are sometimes criticized that we are stodgy people who can be more concerned about rituals and prerogatives than science or the public welfare. It would be unwise to do anything that might strengthen that criticism.
If the President wishes to be the center of attention and seeks to change past practices to enable him to be more so, it is not our job, as scientists or professionals, to challenge it. Our job is to make sure that whatever is released, whenever and by whomever, has reliability and integrity. It is in general for the public to decide whether attention-seeking behavior is consistent with the public interest or not, and we as scientists don’t really have any more to say on the matter than the public generally.
The last thing I would want to do would be to contribute to creating an impression that what’s really going on is we have competing desires, don’t like the spotlight taken away from us, and are cloaking these parochial desires in the garb of scientific integrity. Conduct which could have the effect of contributing to such an impression about us could significantly damage our claim to represent scientific integrity and the public interest. And we are living in a society where it is in many people’s interest to seek out any apparent flaw in us they can find.
For this reason, I believe that before speaking out on an issue as a profession, we need to be very, very confident that (a) it involves grave jeopardy to matters central to our mission, and (b) the risk of us being misconstrued otherwise is very low.
I don’t believe this issue meets either of these two criteria.
Jonathan Siegel
Associate Director Clinical Statistics
Sent from my iPhone
Original Message------
I am dismayed to learn that the ASA President and Executive Director sent a letter to the Washington Post on behalf of our organization. Several thoughts:
1. The hyperbole in the letter is unwarranted ("breach of trust", "disservice to our nation", "alarm", "protect the integrity of government statistics"). What happened? President Trump tweeted one hour before the official release of the employment numbers saying "Looking forward to seeing the employment numbers at 8:30 this morning". Doesn't seem like Armageddon or a massive violation of data integrity to me (although I agree he should not have done it).
2. The ASA was worried about "artificial market disruptions", as though there aren't dozens of firms trying to predict the jobs numbers every month for clients all over the world. Almost all of those predictions in this case were positive. Right now, several major econometric firms are predicting 2Q GDP growth at 4.5% to 4.8%. If Trump tweeted a similar vague positive message, would you be able to read his mind? Would you believe the growth to be 4.2% (he's trying to put a positive spin on numbers which though excellent fall short of predictions), 4.6% (he's just saying what everyone knows/assumes) or 5.2% (the actual is much better than the consensus forecast)? Would you like to bet on the information? More to the point, since EVERYONE gets the tweet, no one will have an informational advantage, will they? Every news story, every rumor, every guest who predicts something, every FOMC report, etc. would cause an "artificial market disruption" in the ASA's view. The markets absorb millions of pieces of data every day, many from sources claiming to have private information.
4. This is not about "data integrity' or any other red herring. There is no statistical issue involved at all. It is blatant partisan politics thinly disguised behind some ethical or methodological concern. Having written to the Washington Post and put the ASA bias on display for everyone to see, we label ourselves as a political organization. Further, by the hyperbole, the ASA discredits itself as just another partisan lobbying group. Very disheartening to someone who wants to belong to an organization respected for its professionalism. The business news media covered this "story" - the ASA should not be involved.
5. This is a good use of the ASA's resources. If you want to use our unique talents and perspectives on public policy issues, I can give you a dozen issues to analyze off the top of my head with far more long lasting import than a Trump tweet.
6. Finally, I wish that the ASA would not get involved in politics and pretend to speak for me or other members. If I want a political organization to speak on my behalf, there are dozens I can join and fund. The ASA should concern itself with statistical issues.
Terry G. Meyer, Ph.D.