NBC phone poll is a disgrace

An NBC poll that was released on Monday received lots of attention but no meaningful analysis. The headline and initial sentence that first led me to this “news” read:

Americans think economy is rigged to favor rich, poll finds

U.S. voters are overwhelmingly convinced that the economy is structured to favor the rich.”

With this ringing endorsement of the Occupy Wall Street protests, I was intrigued. What I found was quite disappointing.  The headline would have been more accurate as “Biased sample of Americans provides an emotional reaction to unsupported allegations”.

The NBC poll was based on phone interviews of 1000 adults. But, the survey participants were not random or representative of the American public.  The survey instructed that the participants be selected as follows:

 Interviews: 1000 adults, including 200 cell phone only respondents. African-American Oversample: Additional interviews were conducted to achieve a total of 400 African-American adult interviews.”

Since the sample is not representative of Americans overall, the results cannot be described as reflecting what Americans overall believe.

Putting aside the sample’s bias, the underlying question is useless as a survey instrument. The question asked whether the respondent agreed that:

 The current economic structure of the country is out of balance and favors a very small proportion of the rich over the rest of the country. America needs to reduce the power of major banks and corporations and demand greater accountability and transparency. The government should not provide financial aid to corporations and should not provide tax breaks to the rich.”

The question was presented as a single take-it-or-leave-it proposition for which the respondent needed to agree or disagree. There was no attempt made to separate the numerous issues that are bundled in this long question. There was no additional explanation as to what the question was attempting to address. There was no lead-in or follow-up question(s) to explore what the respondent knew about these topics. So, the respondent is left to emotionally react without considering foundational questions such as the following:

  1. What is “balanced” mean in the context of this question?
  2. What is the definition of rich?
  3. Are we asking about a small proportion of the rich or a small proportion of the population?
  4. What banking powers are we talking about?
  5. What corporate powers are we talking about?
  6. What greater accountability is desired by an “agree” answer?
  7. What greater transparency is desired by an “agree” answer?
  8. What financial aid are we talking about?
  9. What tax breaks are we talking about?

Even if (i) the sample had not been purposefully biased, and (ii) each respondent knew want was being asked, the question is hopelessly compound. What is being communicated when one responds that that “agree”? Are they agreeing that:

  1. The current economic structure of the country is out of balance, OR
  2. A very small proportion of the population is rich? OR
  3. The current economic structure favors those with substantially more money? (I would hope so! If not, what would be the purpose or benefit of having money?), OR
  4. America needs to reduce the power of major banks and corporations?, OR
  5. Corporations should have greater accountability and transparency in their governance?, OR
  6. Corporations should have greater accountability and transparency in their financial reporting?, OR
  7. Corporations should have greater accountability and transparency in something else?, OR
  8. The government should not provide financial aid to corporations? (And), OR
  9. The government should not provide tax breaks to the rich? (And what tax breaks are we even asking about?), OR
  10. Any one of dozens of possible combinations of the above?

Without a more narrow question that has a complete context, one can only guess what each respondent is thinking when responding that they agree.  Most likley, the various respondents agreed with different parts of the compound question.  Aggregating “agree” responses in this situation does not propertly allow one to report a large agreement percentage.

I am a particularly perplexed about the tax breaks for the rich that this survey attempts to address. Since I also am suspicious of hand-outs, I also am worried about those “rich” who the question contends get a free ride. Perhaps I would have also hurriedly agreed with the question quoted above, so I could return to whatever I was doing when the phone rang.

But my desire to get off the phone and back to my life does not tell anyone about the presumed tax breaks that these “rich” receive. Sure, there are a very few people who have invested well and thereby made good use of the favorable tax rates that apply to dividends and capital gains. Warren Buffet has made a big deal about his situation (without actually backing up his claims with his tax returns).  But how many Warren Buffets are there? There have also been those entrepreneurs who have not paid tax because they actually lost money on business activities, and offset those real losses against other income. But allowing real losses to be offset against gains hardly seems unfair.

However, these tax provisions are not limited to the “rich”, who instead are being taxed using the same rules that affect others. Instead, the rich (who actually would be much better described as the mass affluent) already are denied the vast majority of tax breaks because the majority of these Internal Revenue Code benefits are phased out and eliminated when the taxpayer reports greater income. Importantly, income tax rates are decidedly progressive, so the “rich” already pay considerably more income taxes than others.

NBC should be ashamed to publish dribble like this. The results from this poll do nothing to make us more intelligent.  Instead this “news” incites readers to form conclusions that have no real basis.


Permanent link to this article: http://betweenthenumbers.net/2011/11/nbc-phone-poll-is-a-disgrace/

Leave a Reply

Your email address will not be published.