Statistics for Future Presidents

By Steve Pierson posted 01-11-2012 08:34

I recently started reading Richard Muller's book, Physics for Future Presidents, and it made wonder what statisticians would include in a book of statistics for future presidents. Please let me know your thoughts in the comments space of this blog entry or by emailing me:

I'm aware of lists for the most important statistical skills for statisticians (e.g., The 5 Most Critical Statistical Concepts), K-12 students, non-science majors, and professions such as journalists (e.g., News and Numbers by Victor Cohn) and clinicians. I'm eager to hear what statistics you’d recommend for future presidents (or policymakers more generally) and how it would compare with the recommended statistical skills/concepts for others.

Muller's book organizes his book into a discussion of five topics: terrorism; energy; nuclear weapons, power and waste; space; and global warming. Would a Statistics-for-Future-Presidents book also be organized by such topic areas or would be organized into more statistical thinking categories (interpreting data, assessing study/report conclusions and survey results, decisionmaking in presence of uncertainty, assessing/managing risk, …) or some combination of these?


05-21-2012 07:49

Given how politicians justified their recent votes to defund the American Community Survey, we should better educate them about how statisticians obtain the information that they (hopefully) use in decision-making.
For example, let them know why random and mandatory surveys are essential to getting good information. Here's a Representative bashing the ACS, from the NY Times article linked above: "'We’re spending $70 per person to fill this out. That’s just not cost effective,' [Webster] continued, 'especially since in the end this is not a scientific survey. It’s a random survey.'"

03-21-2012 12:36

I like the idea and the discussion, but I think that instead taking a disciplinary approach we should be fostering analytical thinking. Statistical skills enable a President to deal with quantified information, but many kinds of information come into play for any given decision.

01-25-2012 17:39

I agree with Stephen, for decision makers with limited time on their hands, the book should focus on areas of worry/concern and what they need to know about statistics with regards to that subject area. Ideally it shouldn't be advocating a policy; focusing on how to tell if they're being fed junk or not.
Some of the other suggestions are good for policy wonks behind the scenes, or for bureaucrats, but I think the first idea would have the most widespread audience.

01-25-2012 00:08

Wrong, wrong, wrong! Richard Muller has the right idea. If you want the President (or any non statistician) to read the book it has to grab them at an important nexus of what they worry about and statistics. The book should have chapters on the most important problems of the day in which statistics plays a big role. Here are a few ideas: Global Warming, Economics, Health Care Cost Projections, Medical and Pharmaceutical Trials, Can Statisticians be Trusted?

01-18-2012 12:21

In most introductory statistics classes here at NPS, a number of instructors use something like the following on the first day as course goals. I think some variant of it, geared toward policy makers, would work well in the introductory text.
At the end of the course you should be able to…
- Apply basic statistical methods
- Recognize more advanced statistical techniques
- Distinguish good statistical practice from bad
- Know when to call in statistical experts

01-17-2012 13:23

My suggestions:
1. Decision-making under uncertainty, e.g. decision trees (the OR kind, not the machine learning kind)
2. Bayesian approaches to environmental decisions
3. The hazards of drawing conclusions from observational data (Stan Young's article in Significance, Sept. 2011)
Peter Bruce,

01-11-2012 12:05

I think that this is a great idea. I like an organization by statistical thinking areas with policy-relevant examples that frame the discussion. I suggest including areas like
-- the importance of question wording and context, study design for causal claims and survey statistics
-- distinguishing pattern from random chance (importance of measures of uncertainty)
-- decision making in the presence of uncertainty (decision theory ideas, probabilistic considerations of risk)
-- how wise statistical analysis informs debate (basically, fight the perception of lies, damn lies,and statistics)
-- maybe a chapter on what statisticians can do beyond the intro stat level.
Jerry Reiter
Duke University