Random Thoughts (Scary Ones)

May 21st, 2012 at 2:02 pm

OK.  Sometimes in this town you’ve really got to pinch yourself to see if you’re having a nightmare or did an elected official really say what I thought he said?

So it is with Congressman Daniel Webster’s (R-FL) defense of his bill to eliminate the American Community Survey, today’s version of the Census data gathering that’s been a core input to what we know about America for well over a century.

Rep Webster argued that the survey is a waste of money because it’s random: “We’re spending $70 per person to fill this out. That’s just not cost effective,” he continued, “especially since in the end this is not a scientific survey. It’s a random survey.”

O-M-G.

Random sampling is what makes such a survey scientific.  That’s the way you ensure that your sample is representative.  If you ask folks going into a baseball stadium what’s their favorite sport, your sample is non-random and unrepresentative.  In fact, random sampling is a much cheaper and more efficient way to gather information because with large enough samples, you can get representative information without surveying everyone.

As my CBPP colleague Arloc Sherman points out, there’s a lot riding on making sure this legislation is defeated and not for liberal or conservative reasons, but so businesses, governments at all levels, researchers, and so on can analyze the granular information that a RANDOM survey like the ACS provides.

I get that we’re not exactly in fact-driven mode these days when it comes to the national budget debate, but I think any objective observer would admit that this level of either obfuscation or ignorance goes too far.

As Arloc pointed out to me, you can only assume that the same care and attention that went into this comment also went into consideration of the consequences of eliminating the ACS — for all the businesses (such as Target) that rely on the data better serve their customers, and for communities who rely on the data to plan their school needs, receive their appropriate share of federal funds related to highways or medical care, or in dozens of other areas.

Print Friendly

6 comments in reply to "Random Thoughts (Scary Ones)"

  1. Misaki says:

    Relevant:
    http://woweuarchive.yg.com/thread.html?topicId=1151565148&sid=1
    http://wowarchive.yg.com/thread.html?topicId=2215432134&sid=1

    Poll Results For: “.999~ = 1 ?”

    Yes and I’ll explain at great pain to everyone else why 7191 votes
    48.25%
    No and I’ll explain at great pain to everyone else why 5582 votes
    37.46%
    Huh!? 2130 votes
    14.29%

    Total Votes: 14903

    .999~ = 1 1785 votes
    51.62%
    .999~ 1 1174 votes
    33.95%
    ??? 499 votes
    14.43%

    Total Votes: 3458

    When I was in high school, my math teacher insisted that .999~ was not equal to 1. The next day he corrected himself, after having asked on an online math forum about it. (He also taught statistics, coincidentally)


  2. D. C. Sessions says:

    Reminds me of a Pennsylvania Senator a while ago who wanted to restrict Weather Service data access to protect the profits of the commercial weather services.

    In this case, there’s lots of money to be made by private polling organizations who could charge Target and others good money for the same data that Uncle Sam gives away.


  3. urban legend says:

    How do you hand this clown his deserved humiliation? How about sponsoring a state referendum that would make him change his name so he can no longer disgrace a famous name in American history?


  4. Marcus Sabathil says:

    In Canada, our current “conservative” Government has been slashing and burning our world famous Stascan. Their latest budget puts HALF of Stascan jobs at risk. Stephen Harper also recently killed the long form census for ideological reasons. My guess is that the truth is inconvenient. Private, corporate info is much simpler to control. Take drug studies, for instance. If the data does not support your desired conclusions, don’t publish.


  5. David A. Spitzley says:

    I’m assuming that this is a further evolution of the thinking that resulted in Congress attacking the use of random sampling in the Census process, due to the fact that it tended to increase the numbers of minorities, as well as boost population counts in Democratic-leaning states. They’ve basically gone from “random sampling is being used in ways which undermine the GOP politically” to “random sampling is not scientifically valid”, and don’t realize why that’s incorrect.


Leave a Reply

Your email address will not be published.

Current month ye@r day *