Bryan Caplan  

Survey Pre-Test: Who in Government Has Power Over What?

Rothbard on Szasz... Greg Mankiw Pumps for a Higher...

A few months ago, readers of this blog gave helpful comments on the first draft of the Caplan-Somin survey about who in government has power over what.

Now we're ready to pre-test.

If you've got the time, please take this survey at, then tell us how we could improve it. Feel free to post comments on the blog, or just use the mail form at the end of the survey.

Much oblige!

P.S. Ideally, this survey would have a bunch of short pages instead of one long page, but I'm not a good enough Perl programmer to make this happen. Any Perl gurus out there who could quickly re-write this as a multi-page survey? Lunch is on me if you do. :-)

Comments and Sharing

TRACKBACKS (3 to date)
TrackBack URL:
The author at Philosophy Me in a related article titled Political Study at EconLog writes:
    The fine folks over at EconLog are looking for help in pre-testing a survey regarding "Who in Government Has Power Over What?" by Bryan Caplan and Ilya Somin. The survey itself doesn't take very long, and poses some very interesting questions about t... [Tracked on October 21, 2006 7:52 PM]
COMMENTS (18 to date)
mobile writes:

Is #4 a trick question? Or is someone trying to peel off my tin foil hat and tell me that the Federal Reserve Board is a part of the federal government?

Doug writes:

I thought the section where you score different parts of government 0-10 vs. different questions was poorly implemented. Am I supposed to be assigning a total of 10 across each row - i.e. Congress has 20% of the power? Or am I free to assign lots of 10s in the same row?

JMB writes:

Just a couple of things that I found unclear from the instructions given:

I don't get the question about whether various people/groups have more power to improve things or to make things worse. My libertarian gut instinct is to say that their primary power is to make things worse, and that we'd be better off if they all sat on their hands. But, since (from a libertarian perspective) most things are already screwed up, in many instances they could all make things better by first rolling back previous regulations and *then* sitting on their hands. So I couldn't come up with a coherent answer for any of those.

I also found myself really fighting the urge to over-think Part 3. It's hard to say that any of them have *no* power at all to affect those things, since theoretically if they *really* wanted to most of those branches could screw things up enough to negatively affect just about anything. Like, if the Federal Reserve really wanted to, it could wreck the economy badly enough to bring down the average life expectancy, but in the normal course of things they have essentially no power to affect life expectancy. So which of those are you looking for--the absolute limits of their power under the law, or their power in the normal course of things?

In the same question, the DoD and the EPA are somewhat confounded with the president--DoD and the EPA can, to a large extent, only get away with what the president will let them get away with. So, for example, the EPA could promulgate regulations that will have a negative effect on the economy in 2 or 20 years, but only if the president lets them. So in that case should one ascribe responsibility to the EPA (who actually did the act) or the president (who at minimum did not stop the act, and therefore in some sense has ultimate responsibility for it)?

Rue Des Quatre Vents writes:

I agree with Doug. It's not clear whether we should be thinking of federal power in zero-sum terms. Also it's unclear until the end whether we should be thinking in terms of the power to make things worse or the power to make them better. For example, I don't believe the President has the most power in the federal government to improve the economy, but by the push of a button, he sure has a lot of power to destabilize it.

FC writes:

On part 2, question 6, I think the phrase "strike down" is associated in common usage solely with the courts, especially the US Supreme Court. You might want to reword the question so people won't unnecessarily exclude Congress and the President from their considerations.

kenf writes:

You can probably post the PERL script on and get a rewrite done for $10 or something.

Mike Linksvayer writes:

I hate multi-page anything. Why make users click more than necessary?

I'm eager to see the results! The various branches and levels have little power over anything, except to make things worse. :)

Carl Marks writes:

Part One
2. Define war. Technically the last time we were at war was with romania (1943?). But i think you mean more about having troops in combat.

Part Two.
6. Good one, i suspect many will wrongly choose the courts

Part 3 should not be numbers, rather a "not at all, a little, total control, etc." metric

Dr. T writes:

1: "Education policy" is too broad. Do you mean K-12 education, higher education, both?

4: I can't think of a correct answer. Monetary policy can be affected by Congressional budgets, presidential actions, Federal Reserve Board actions, and (via money supply) the Treasury Dept.

7: Technically, Congress decides (via the budget) how tax money is spent, but the executive branch can unilaterally alter it. For example, Congress could authorize a Federal Dept. of Economics and allocate a billion dollars to construction of a building. If the executive branch drags its feet, nothing will ever get spent.

8: Which elections? The federal government has much control over federal elections but almost none over village mayor elections. Since there are many more local elections than federal, I would answer state and local. However, federal elections draw many more voters, so one could reasonably conclude that federal power is greater.

Part 3: The degree of income inequality. What do you mean by that? Janitor wages vs. CEO compensation? Average income of working poor vs. upper 5%. Me vs. Bill Gates?

Overall, I dislike the word power in the questions. I prefer two words: control and influence. For example, Congress controls the budget but the President has tremendous influence on the budget. (So much so that in modern times the Presidents send proposed budgets to Congress.) I realize that your grid questions attempt to address my complaint, but even there the word power is used when affect or influence might be better.

Eric Crampton writes:

The second write-in question -- hard to tell the right answer here. Only Congress has authority to declare war, but they seem to have delegated such authority to the President of late.

powerless writes:

On part four, I think that all the agencies mentioned have more power to make things worse than to improve things.

But I think that is a general feature of the human condition. For example, as a parent, I could, if I set my heart to it, make my children miserable. I try my best to make them happy, but I am only so good at it -- and so many factors are out of my control.

Brad Hutchings writes:

Any way you could make it shorter? Or break it up into multiple surveys to be run one at a time on an unsuspecting public? When I got to filling in the 1-10 in that matrix, I was thinking, "this is more work than an upper division government midterm". I guess the Internet is short attention span theatre. I plead guilty to being in that audience. Hey, could that actually bias your results?

Rochelle writes:

What, no libertarian demarcation under political party/political beliefs?
What's up with that?

sa writes:

is the survey only for US citizens?

Tony writes:

When you ask the question about power do you mean power relative to the other parts of government you listed or relative to some absolute?

I don't know if breaking down all of the agencies/departments is a good idea. There are too many that should also be included such as labor (osha davis/bacon) transportation (faa), justice (ada), agriculture, interior (land use), and education. Would it be better to lump all regulator agencies into one basket?

Horatio writes:

How do you define liberal? I consider myself very liberal and clicked that box in the survey. However, what most Americans consider liberal is more aptly described as socialist.

Christina writes:

I agree with those who have said the numerical ratings section is unclear. If we're indended to rate so that the line totals to 10, that's very different from evaluating each one independently.

It's certainly an interesting exercise in what we understand about governmental power and influence. I look forward to reading about your findings!

Pretinieks writes:

Many people need some "warming up", so any survey must start with at least some easier introductory questions. In my country, questions like "did you vote in last election?" and "in general, do you have much interest in politics?" are usually placed at the start even if pollsters have no interest in the answers :)

Open-ended questions are not easy, and placing them first is a sure way to make people angry because "these guys demand the impossible".

Perhaps you could place Part 4 first, or invent some new "chatting" questions.

U. (Riga, Latvia)

Comments for this entry have been closed
Return to top