Candidates
cy scholars, or perhaps for the Heritage interns
libertarians people who are really conservative--
the word "libertarian" is unfamiliar or undesir-
we surveyed interns and analysts at the Heritage
and political
able. If the latter is the case, we should expect to
Foundation, most of whom we assumed would
strategists may
be well-informed conservatives.64
get more people to self-identify as conservative
find a large group
when they are really more libertarian.
Heritage respondents self-identified as
Do more questions or a different set of
follows:
of homeless
questions do a better job? No, we found that
voters looking
adding more questions on specific topics
14% libertarian
such as the Patriot Act, homosexuality, or
9% moderate
for a home.
family values did not add much to the relia-
55% conservative
bility of the results. Thus we continued to use
21% very conservative
the broader questions to screen respondents
2% not sure
to the national polls. On balance, we believe
· Using
these results confirm the validity of our
the libertarian screening ques-
results from Gallup, Pew, and ANES.
tions from ANES, 31 of 106 (or 29 per-
cent of) Heritage respondents would be
"libertarian." Those include 10 of 15 (or
Notes
66 percent of) self-identified libertari-
ans, 5 of 22 self-identified very conser-
The authors wish to thank Isaac Freilich Jones for
vative, 14 of 57 conservatives, and 2 of 9
indispensable data crunching.
moderates.
1. Raymond E. Wolfinger et al., The Myth of the
·
Using the libertarian screening questions
Independent Voter (Berkeley: University of California
from Gallup, 28 of 106 (or 26 percent of)
Press, 1992).
respondents would be "libertarian."
2. Joshua Green, "In Search of the Elusive Swing
Those include 14 of 15 self-identified lib-
Voter," Atlantic, JanuaryFebruary 2004.
ertarians, 3 of 22 very conservative, 7 of
57 conservatives, 3 of 9 moderates, and 1
3. "Karl Rove--The Architect," Frontline, PBS,
of 2 "not sure."
http://www.pbs.org/wgbh/pages/frontline/show
·
s/architect/interviews/.
Using the libertarian screen question
from Pew, 36 of 106 (or 34 percent of)
4. William Crotty, John S. Jackson III, and Melissa
respondents would be "libertarian."
Miller, "Political Activists over Time: `Working
Those include 14 of 15 self-identified
Elites' in the Party System," in Comparative Political
Parties and Party Elites: Essays in Honor of Samuel J.
libertarians, 3 of 22 very conservative, 15
Eldersveld, ed. Birol A. Yesilada (Ann Arbor: Univer-
of 57 conservatives, 3 of 9 moderates,
sity of Michigan Press, 2002), http://www.csd.n
and 1 of 2 not sure.
eu.edu/Political_Activists_Over_Time.html.
5. William S. Maddox and Stuart A. Lilie, Beyond
Our three-question screens are slightly under-
Liberal and Conservative (Washington: Cato Institute,
inclusive from the Cato test, and slightly over-
1984), pp. 2829.
inclusive from the Heritage test. On net, we col-
lect a few more self-identified conservatives
6. Stuart A. Lilie and William S. Maddox, "An
Alternative Analysis of Mass Belief Systems:
than we lose self-identified libertarians. Among
Liberal, Conservative, Populist, and Libertarian,"
the data sets, Gallup questions seem to work
Cato Institute Policy Analysis no. 3, August 24,
best. ANES is slightly more overinclusive. Pew is
1981, http://www.cato.org/pub_display.php?pub
slightly more overinclusive than that.
_id=869&full=1.
Among the Heritage Foundation respon-
7. Michael Barone and Grant Ujifusa, The Almanac
dents, the screens identify almost all libertari-
of American Politics 1982 (Washington: Barone &
ans, exclude almost all "very conservative"
Company, 1981), p. lxiii.
respondents, but include some conservatives.
8. Richard E. Cohen, "Rating Congress--A Guide
Perhaps this is a definitional issue for the poli-
23