Self-identification By Party And Political Philosophy

I have been meaning to get around to this since when the last Quinnipiac national poll came out earlier this month. In the crosstabs for the release, there were breakouts on many questions by party and/or by political philosophy.

My impression, gleaned from many discussions and from reading many news articles is that people frequently use “moderate” as a political philosophy interchangeably with “independent” as a party self-identification. If the overwhelming majority of moderates self-identify as independent, and vice-versa, then this is quite valid. But is it? And regardless of that question, I think it would be interesting to know how these self-identifications relate with each other.

Fortunately, I do not have to guess (much) as to this. While one cannot determine the approximate crosstabs for this from the official poll release linked above, the good people at Quinnipiac were willing to run crosstabs for this upon my request.

Before I get to the below, let me qualify my commentary by saying that I am well aware that this is one survey, and treating the numbers from it as absolutely correct is not my intention. I just do not want to continually be writing something along the lines of “according to this survey’s results” and “within this survey’s margin of error.” Also, Quinnipiac supplied me the margins of error for each crosstab; rather than reproducing them all, I’ll just say that they range from +/-3% to +/-4%. (Side note: they provided them to me to the second decimal point. However, due to the fact that the results were reported as whole integers, that is how I choose to represent the MoE. Significant digits!)

First, let me present the data they shared with me. First, here is the breakout of philosophical self-identification by party self-identification. Read this from top to bottom, meaning for all self-identified Republicans, 3% say they are liberal, 22% say they are moderate, etc.

	Tot	Rep	Dem	Ind
Lib	20%	3%	44%	15%
Mod	36%	22%	40%	46%
Cons	40%	74%	13%	36%
DK/NA	4%	1%	3%	4%

I am struck by how Republicans overwhelmingly call themselves conservative while Democrats are nearly as likely to call themselves moderate as liberal. Also, Democrats are roughly 3 times as likely to call themselves conservative as Republicans are to call themselves liberal.

Second, here is the breakout of party self-identification by philosophic self-identification. Again, read this from top to bottom.

	Tot	Lib	Mod	Con
Rep	26%	3%	16%	49%
Dem	31%	67%	34%	10%
Ind	34%	24%	43%	30%
Other	7%	5%	5%	8%
DK/NA	2%	%	2%	3%

From the above, a little math can give us rough approximations of the percentages for each combination. Granted, I cannot account for rounding error here, so by no means should one consider these to be absolute (both for this reason and for the caveats specified above). However, they should be close enough for government work in terms of this survey.

(Note above that Quinnipiac did not give an “other” option on the philosophy self-identification question. This makes sense as they are breaking the entire spectrum down into three buckets. There really is not an “other” for this purpose. Meanwhile, on the party self-identification question they did, which again makes sense as there are parties beyond Republicans and Democrats, so the Independent label is not a catch-all for those who gave an answer. For my purposes, I have rolled “don’t know/didn’t answer” into “other”, mainly because both the “other” and “don’t know/didn’t answer” buckets provide no information as to which breakout most closely applies.)

Lib/Rep	1%	Mod/Rep	6%	Con/Rep	19%	Other/Rep	0%
Lib/Dem	14%	Mod/Dem	12%	Con/Dem	4%	Other/Dem	1%
Lib/Ind	5%	Mod/Ind	16%	Con/Ind	12%	Other/Ind	2%
Lib/Other	1%	Mod/Other	2%	Con/Other	4%	Other/Other	1%

Putting these results into an ordered list:

Philosophy/Party	Pct
Con/Rep	19%
Mod/Ind	16%
Lib/Dem	14%
Mod/Dem	12%
Con/Ind	12%
Mod/Rep	6%
Lib/Ind	5%
Con/Other	4%
Con/Dem	4%
Mod/Other	2%
Other/Ind	2%
Other/Other	1%
Other/Dem	1%
Lib/Other	1%
Lib/Rep	1%
Other/Rep	0%

There are many different potential ways to interpret this. I have my own, but my purpose in posting this is not to advocate any particular political position; as such I will save that for a possible later post. Instead, my purpose in posting this is to provide context for interpreting poll numbers. I want to thank Quinnipiac for providing me details that they had not released as part of their publication of this survey, particularly April Radocchio, Tim Malloy, and whoever in their tabulations department who took the time to fulfill this request. Thank you very much!