How (Ideologically) Different is a Rur’l Republican Texan from an Urban Democratic New Yawker?

How polarized is the American electorate? According to Fiorina et al’s “Culture War? The Myth of a Polarized America” not very. Fiorina et al. do a wonderful job of laying out the claim that it is the elites that are polarized, not the electorate. So how ideologically different is a Republican rur’l Texan from an urban Democratic New Yawker? Here I plot the mean ideology (measured on a -1 to 1 scale with -1 being extremely liberal and 1 extremely conservative) for the US, rural Republican Texans and urban Democratic New Yorkers.

D-NY.R-TX.png

My co-blogger John Sides noted that “the liberal fair-trade coffee-drinking Per Se-dining New Yorker is closer to the US median than the conservative truck-drivin’ Toby Keith-lovin’ Texan.” This simple analysis does seem to call into question the perceived notion that red America is more authentically “American” than blue America, but we’ll explore this in more detail in some further posts.

9 Responses to How (Ideologically) Different is a Rur’l Republican Texan from an Urban Democratic New Yawker?

  1. Kevin November 27, 2007 at 10:00 am #

    Where does this data come from?

  2. John Jenkins November 27, 2007 at 10:20 am #

    Doesn’t the fact that there are more of the New Yorkers virtually ensure that if you plot the mean (as opposed to median) ideology, you will always find the New Yorker closer to the average than the Texan because the average is weighted so heavily toward the New Yorker?

  3. Sk November 27, 2007 at 11:27 am #

    You should run this by some social scientists. Because, as is, its meaningless.
    Defining liberal as ‘-1’ and conservative as ‘+1’ is a meaningless statistic-we don’t know if ‘-.5’ or ‘+.5’ is really really close or really really far from the average (is the scale linear, log, what?).
    Your graph suggests 0 is the mean. Your text suggests the 0 is the median. Which is it?
    The US average is greater than 0? What does ‘average’ mean if not the average of the sample-or in these graph terms, ‘0’?

    ‘This simple analysis does seem to call into question the perceived notion that red America is more authentically “American” than blue America, but we’ll explore this in more detail in some further posts’ No it doesn’t. It calls into question the notion that a rural Texan is closer to the mean (or median?) than an urban New Yorker. Whether a rural Texan is identical with ‘red America’ (and whether an urban New Yorker is identical with ‘blue America’) isn’t addressed.
    As the previous poster already noticed, if ‘authentic Americanism’ is defined as ‘what’s most popular in America,’ then the urban population, which is greater than the rural population, will statistically win that debate. I doubt if anyone who believes in ‘authentic American’ defines it as ‘what is most popular.’

    sk

  4. VentrueCapital November 27, 2007 at 12:26 pm #

    I agree with Sk’s comments — and also am dubious of the validity of any unidimensional ideological scale. I have enormous respect for Fiorina, and I still think his work is flawed because he doesn’t have a valid — or even unambiguous — definition of “liberal” vs. “conservative.”

  5. David K. Park November 27, 2007 at 12:58 pm #

    Kevin,

    The data comes from the 2000 National Annenberg Election Studies. I plan on looking at the 2004 Annenberg study to see if the numbers are comparable. I’ll list the data source on future posts.

  6. David K. Park November 27, 2007 at 1:06 pm #

    John,

    Here are the sample sizes for the analysis:

    US – 58,383
    Urban Dem NY – 672
    Rural Rep TX – 199

    So you can see that Urban Dem NY’ers make up a relatively small sample size compared to the overall sample.

  7. Tom Holbrook November 27, 2007 at 5:27 pm #

    Viewed from a different perspective (Rabinowitz et al.), though, the rural Texan is on the “same side” as the median voter.

    BTW, good to see your blog up and running.

  8. David K. Park November 27, 2007 at 8:14 pm #

    SK,

    These are means, and not medians. I should have corrected John’s quote. Each respondent is asked to place themselves on a scale from -1 to 1 (technically, it’s 1 to 5, but I recoded the variable so it goes from -1 to 1) where

    -1 Very Liberal
    -.5 Liberal
    0 Moderate
    .5 Conservative
    1 Extremely Conservative

    I then estimated the mean ideology for those who identified themselves as Democrat, living in an urban area, and in NY; and a Republican living in a rural area, and in TX. So the mean of all the respondents (here I’m calling it the US) is 0.082, slightly right of moderate.

    You’re right that Red America does not equal Rural Republican Texan and Blue America does not equal Urban Democratic NY but I think for many people these two areas typifies Red and Blue America. I’ll estimate Red, Blue and Purple ideology means and add that to the graph as well. Thanks for the suggestion.

    But the larger question that you raise, what is the meaning of self-reported ideology is a good one and worthy of a post. Recent political science research have examined questions such as, “Is there any correlation between self-reported ideology and issue stances, as well as vote choice?” and “Can we scale self-reported ideology measures and issue stances they can consistent meaning across individuals, states, etc?”

  9. David K. Park November 28, 2007 at 10:13 pm #

    Tom,

    Thanks. Good point about directional versus proximity models. I’m assuming referring to the Rabinowitz and MacDonald (1989) APSR article “A Directional Theory of Issue Voting?”