Hi Everyone:

Given that there's currently a government shutdown of non-essential government employees, and as a former government non-essential employee, I have a question to ask: Did I do anything worthwhile in my six years of working for the US Government?

To give some background, I left my home at Virginia Tech in November of 1975 and went to work for the United States Census Bureau in Suitland, Maryland (which is right outside of Wshington DC) as a Mathematical Statistician. My first three-and a-half years I worked in Statistical Methods Division with the vast majority of my work being on something called The Annual Housing Survey which (I believe) still exists today and is now called the American Housing Survey (AHS). Anyway, after three-and-a-half years of this I was an expert on things like how many rooms were in the typical house, how mobile homes were difficult to find, and what the differences were between urban and rural housing.

https://www.census.gov/programs-surveys/ahs.html

My next two-and-a-half years were spent in Laguna Niguel CA at one of the three processing centers doing Quality Control work for the 1980 Census. This included a period of over six months where I worked a minimum of 10 hours a day, didn't have a day off, and was not paid for most of this overtime. I also had as much as 700 people working for me. And even though I was still a mathematical statistician, I think most of my time was spent dealing with bad clerks who didn't want to do their job.

As a result of my six years with Census, today I refuse to participate in any surveys and believe that the number of questions on the decennial Census should be reduced. Also, to show I actually did a few things during my time at Census, in August of 1978 I presented a paper which I wrote at The American Statistical Convention titled "The Advantages and Disadvantages of Rotating the Annual Housing Survey, National Sample From a Nonresponse Point of View" and below you will find a link to a Statistical Policy Working Paper where my name is mentioned:

https://nces.ed.gov/FCSM/pdf/spwp3.pdf

All comments are welcome.

Best wishes,
Mason

PS: As an aside, one worthwhile thing I did learn was that it's best to design statistical surveys so that they are "self-weighting." When I got into gambling, I immediately understood that you need to do the opposite if you want to be successful at games like poker. Those of you who have read my Gambling Theory book will be familiar with what I call "Non-Self-Weighting Strategies." Well, my years at Census is where this came from.

Also, my best friend from graduate school and my years at Census was a man named John H. Thompson who became the 24th Director of the U.S. Census Bureau from 2013 into 2017. Ironically, the last time we talked was during the government shut-down under President Obama. He actually spoke to me from his office at the Census Bureau and told me he was one of the only people there. Here's his bio, and notice that his degrees are exactly the same as mine and his early years at Census he also worked in Statistical Methods Division:

https://en.wikipedia.org/wiki/John_H._Thompson

Why did your work there make you not want to participate in any surveys?

Do you participate in the census?

How does random weighting compare with other forms of weighting? Is it maximally bad?

I participate in those surveys that are required by law. For instance, on behalf of Two Plus Two, I recently answered the The Economic Census.

However, one of the things I learned and saw first hand was how intrusive a government survey can be, and as the years went by, I began to feel uncomfortable about it.

Best wishes,
Mason

You need to read my Gambling Theory book. But a quick question: When you play poker, do you play every hand? If you do, then that's part of a self-weighting strategy.

Best wishes,
Mason

Yeah most likely. Honestly one of the most important services the federal government largely gets right is the VERY valuable statistics it compiles. Speaking as someone with a light econ background (just undergrad LOL) those statistics are insanely useful to decision makers of all sorts.

Why do you feel that the questions on the census are too intrusive? What's the downside to the decision makers of the world knowing more about the typical citizen/consumer/worker rather than less? I'm MUCH more concerned about things like facial recognition, cameras everywhere, cell phone metadata tracking, and all of the other various digital leashes that let the powers that be monitor me personally than I am about them knowing how many rooms the average house has. In fact I think data collection of the latter type has mostly made the world a better place.

With the caveat that I know nothing about you save that you co-authored Theory of Poker and help run this website, I’d hazard to guess that your work with the census was exponentially more useful, I.e. productive to society, than your work on gambling since

Here's an excerpt https://www.twoplustwo.com/books/gam...theory-topics/

My immediate reaction to his was "For better or for worse, I bet Facebook, Google, and/or Amazon have more and better data about America and Americans than the federal government."

Only about the America and Americans they can monetize. I doubt they know or care much about the bottom rung of the economic ladder.

Afaik, the census provides highly valuable information to many businesses and planners. As the census is required by the Constitution and necessary for a representative form of government more generally, the question you raise, about whether your work was "worth it" (a question that can be asked about any work, even writing poker books), is whether the census should be (relatively) broad or narrow. That's mainly a question for the academics and businesses that use census information. My general sense is that in terms of cost/benefit the census is likely a pretty good deal. Probably more so that the vast majority of defense spending. But you be you and keep sticking it to man and his desire to compile accurate information about the populace.

+1

Census data is also enormously important to social science researchers. I <3 the census. Thanks Mason.

Mason,

The universe is nearly 14,000,000,000 years old. Your entire life and everything you've ever seen and done and ever will do accounts for at best 0.0000007% of the history of the universe. The Milky Way alone has about 200,000,000 stars, the actions of one sentient being are utterly meaningless even on a global scale let alone a galactic one and our galaxy is only one of the trillions of galaxies in the observable universe which is possibly one of an infinite number of parallel universes in the multiverse.

So no. Probably not.

Need three more zeroes, unless the galactic census is way off.

Actually my question was completely serious, and ought to make perfectly good sense to you. I'm asking what's the analogue in this setting of an entropy maximizing prior.

Probably not true. They aren't asking the same questions or collecting the same data points.

Luckily, in the scheme of things my error matters even less than masons life.

Sounds like Mason is rationalizing a permanent government shutdown as actually a really good thing

Seems like a good justification for mass murder being ok.

Or it's a totally mistaken way of looking at it - I prefer to believe this.

re: census. It depends on who you're doing it for. It's a powerful tool.

This guy sure does love assuming the conclusion.

“I’ve internalized right wing derp about filling out the census being a form of slavery (thereby telling on myself regarding the media I consume), therefore when I worked for the census that was a waste of time. Also I was in charge of 700 people because I’m awesome.”

If you don't invoke the argument that someone else would have eventually done the same thing then you would have to switch your conclusion when speaking of Lise Meitner.

That is actually not true for specifically poker even if it might be true for other games. If everyone was as adept at probability, game theory, psychology, and logic as you need to be to get good at poker, the world would be improved to a greater degree than if governments were all well versed about various statistics of their populations.

But there would be no money in poker; everyone would be solid.

Why does this thread exist?

Are you really looking for the assurance of complete strangers that what you did was a worthwhile endeavor? All of that is in the eye of the beholder. If I think the Census is important, it means dick if you don't think that way.

Personally, I think it's an important tool. But then again, you should know its value better than most.

If you and nobody else suffered as a result of your actions, you did something worthwhile.