ANES logo
                      Memos to the ANES User Community

Original Document Date:  June 30, 1994
Date of this Revision:   June 30, 1994
Filename:                \ftp\announce\newsltr\ltr40694.wp

June 30, 1994

MEMO TO:  ANES User Community and Other Interested Social

FROM:     David C. Leege, Chair, ANES Board of Overseers

RE:       Research and Development Conferences in Preparation
          for the 1996 National Election Study

Over the next six months, ANES anticipates convening three
research and development conferences:

    Conference on Candidate Evaluation
    Conference on Values and Predispositions
    Conference on the Impact of the Presidential Campaign

I am writing to describe our initial plans, to solicit your
comments and suggestions, and to invite your participation in
these activities.

The enclosed memos by Virginia Sapiro and by Larry Bartels and
John Zaller describe a broad range of questions on which we would
like to stimulate discussion and around which we anticipate
organizing these conferences.  These memos are meant to provide a
sampling of the issues that might be addressed.  Please feel free
to suggest other issues that occur to you.

Each conference is designed to be a vehicle for involving the ANES
user community and other interested social scientists in shaping
the design and content of the 1996 National Election Study. 
These activities are quite consciously intended to be occasions
at which social scientists can assess the utility of existing ANES
instrumentation and study designs, and can suggest innovations
and improvements that should be made.

Planning for the 1996 American National Election Study will also include,
among other things, a Pilot Study that will be conducted in 1995.
The 1995 ANES Pilot Study will provide an opportunity for social
scientists who have suggestions for new instrumentation for the
1996 Election Study to test out those ideas.  We will be
circulating a memo this fall inviting your participation in that

Finally, in 1996, ANES will convene a Conference on Congressional
Elections Research designed to do two things:  to assess the
performance of core instrumentation first introduced in the 1978
Election Study and carried in each subsequent Study; and to
identify the next steps that should be taken in the ANES studies
of Congressional elections.  We will also circulate a separate
memo next year inviting your contribution to that gathering. We
anticipate that the recommendations from this Conference will be
tested in the 1997 ANES Pilot Study with an eye towards
implementation in the 1998 midterm Election Study.

Participants for the three Conferences that will be convened this
fall will be chosen on the basis of memoranda or papers submitted
to the Board.  Memos of 2-3 pages in length that respond to one
or more of the questions or issues posed in the enclosed
memoranda would be most helpful.  If a paper is being offered for
presentation at one of the Conferences, the paper does not need
to be submitted at this time, though a brief description of its
content would be appropriate.  The deadline for submission of
your response is August 26, 1994.

Please direct your suggestions and reactions to:

     Research and Development Conferences
     The American National Election Studies
     Center for Political Studies/ISR
     P.O. Box 1248
     Ann Arbor, Michigan 48106-1248
You can also reply by email to:

If we can provide any additional information about the activities
described in these memos, please feel free to contact any member
of the ANES Board of Overseers or one of the Principal

           - Board of Overseers -

David C. Leege, Chair         Larry Bartels
University of Notre Dame      Princeton University
(219) 631-7809                (609) 258-4794

Charles H. Franklin           R. Douglas Rivers
University of Wisconsin       Stanford University
(608) 263-1878                (415) 723-2612

Virginia Sapiro               John Zaller
University of Wisconsin       University of California, Los
(608) 263-1873                (310) 825-4331
 until July 25, 1994)

          - Principal Investigators -

Steven J. Rosenstone          Donald R. Kinder
University of Michigan        University of Michigan
(734) 764-5494                (734) 763-4783   

Warren E. Miller
Arizona State University
(602) 965-3034

June 30, 1994

MEMO TO:       ANES User Community and Other Interested Social

FROM:          Virginia Sapiro

RE:            Conferences to Critically Evaluate Recent
               to "Core" ANES Instrumentation

At the heart of each Election Study are a set of "core" questions
that ANES asks regularly in its studies of the American
electorate. Over the years, through the ANES Pilot Studies,
members of the ANES user community have improved the ANES measures
and have introduced new concepts and new instrumentation that
have been added to "core."

The Board of Overseers has targeted three sets of core items for
scrutiny:  candidate evaluation (introduced in 1980); values and
predispositions (introduced in 1984); and contact between
citizens and candidates for the U.S. House of Representatives
(added to ANES core in 1978).  Each body of instrumentation was
developed in the formative years of the ANES as a national

Over the next year, with support from the National Science
Foundation, the ANES Board of Overseers anticipates convening
several Research and Development Conferences.  ANES will bring
together members of the user community to take a hard look at
its instrumentation:  this includes evaluating the utility of
the concepts and adequacy of the theories; assessing the
performance of the items; and exploring new ways the concepts
might be measured.  At this time, we invite your participation in
two of these conferences:  one on candidate evaluation and
another on values and predispositions.  Next year we will
circulate a memo describing in more detail our plans for the
Conference on Congressional Elections Research.

     Conference on Candidate Evaluation

Presidential candidates succeed or fail in part because of the
qualities of character they present.  Drawing upon theory and
research in social psychology that underscores the importance
of character in interpersonal evaluation, ANES developed and
tested, in a series of ANES Pilot Studies, candidate trait

In various forms, these candidate trait batteries have appeared
on the National Elections Studies conducted since 1980.  An
abundance of evidence shows that Americans judge Presidents and
presidential contenders by their competence, their integrity,
their compassion, and their leadership and that these
assessments substantially influence the choices citizens
eventually make.

A second aspect of candidate-centered voting appears to be
emotional.  Through the ANES Pilot Studies, an inventory of
emotions that candidates evoke was also developed and has been
included in National Election Studies conducted since 1980.  The
evidence demonstrates that emotional reactions that candidates
evoke contribute to the voter's choice, over and above the
effects due to party attachments, policy views, assessments of
the national economic conditions, and judgments of character.

The Board of Overseers seeks your advice on a range of issues:

    * What is our current theoretical understanding of the ways
      in which citizens evaluate candidates for public office?
      What is the current state of knowledge about the cognitive
      and affective components and dynamics of candidate

    * Does the ANES instrumentation do all that it can to support
      the rigorous and fruitful analysis of the ways in which
      citizens evaluate candidates?  How well do the ANES measures
      of candidate perception, traits, affect and other related
      items perform?  What improvements should be made?

    * Are there other approaches that ANES should take in its
      effort to understand the personal ways in which citizens
      assess candidates?

      Conference on Values and Predispositions

Over the last decade, ANES has made a substantial investment in
instrumentation designed to advance our understanding of the
antecedents of public opinion.  For instance, ANES has developed
measures of general predispositions that, in principle, should
influence opinions on a wide range of public policy questions,
including measures of individualism, moral traditionalism,
patriotism, militarism, racial prejudice, limited government,
traditional conceptions of American identity, gender
consciousness, religiosity, and egalitarianism.  These measures,
to varying degrees, have appeared on National Election Studies
since 1984.

In addition to identifying general predispositions that might
influence public opinion across a wide range of issues, ANES has
also gone more deeply into the antecedents of public opinion in
three selected areas:  on race policy carried out as part of the
1986 Study; on foreign policy in the 1988 Study; and on public
reactions to war, a centerpiece of the 1990-91-92 Political
Consequences of the Persian Gulf War Study.

The Board of Overseers would like your advice on a number of

    * What theoretical utility do values and predispositions have
      in theories of public opinion and electoral behavior?  What
      is the state of our knowledge about the role of values and
      predispositions in shaping public opinion, political
      participation, and electoral choice?  What do we have
      yet to learn?

    * What is the utility of the ANES measures of values and
      predispositions?  What do we learn by carrying this
      instrumentation on ANES surveys that cannot be learned
      by measures of issue positions alone?

    * What evidence is there that the measures of values and
      predispositions are largely exogenous to the political
      process, or at least to the ebb and flow of forces
      during a political campaign?

    * How much breadth and depth of coverage of values and
      predisposition is necessary in ANES surveys, and for
      what purposes?  How frequently should these measures be
      included?  On panel studies, do we need to repeat items
      on every wave, or are responses stable enough that
      asking questions on a single wave is sufficient?

    * How well do the various values and predispositions perform
      (e.g. what is the reliability, stability, validity,
      unidimensionality, predictive power), particularly when
      put head to head with other values and predispositions
      or measures of opinion on issues?  Are there particular
      measurement problems that need to be addressed? What are
      the limitations of the existing dimensions and items?  How
      might these measures be improved?  What additional
      dimensions (if any) should be explored?

   *  What are the new ways by which ANES should push forward
      research into the role that values and predispositions play
      in public opinion and electoral behavior?

The Board invites interested social scientists to send us a 2-3
page memo or an already written paper responding to one or more
of these issues.  Please feel free to suggest other issues for
either Conference that occur to you.  If a fresh paper is being
proposed for presentation at the Conference, the paper does not
need to be submitted at this time, though a brief description of
its content would be appropriate.  The deadline for submission of
your response is August 26, 1994.  Please direct your reply to:

     Research and Development Conferences
     The National Election Studies
     Center for Political Studies/ISR
     P.O. Box 1248
     Ann Arbor, Michigan 48106-1248

You can also reply by e-mail to:

June 30, 1994

MEMO TO:       ANES User Community and Other Interested Social

FROM:          Larry Bartels and John Zaller

RE:            Conference on the Impact of the Presidential

Presidential campaigns are the biggest and most engaging of this
country's democratic rituals, but somewhat surprisingly, social
scientists have yet to demonstrate that they make much
difference.  To the extent that scholarship has weighed in, it
has tended to be, at least until recently, on the side of
non-effects.  The state of the economy, whether the country has
been at war, the positions embodied by the candidates, and the
popularity of the current incumbent all affect election outcomes. 
The effect of campaigns on important outcomes, including their
effect on civic education and culture, has only been partially

One possible reason that important effects of presidential
campaigns have not so far been convincingly demonstrated is that
they do not exist.  It would not take an obtuse reading of the
early classics of American voting behavior to arrive at this
conclusion.  The success of models that can predict election
outcomes before the campaign begins likewise lend credence to
this possibility.

Other possibilities, however, are that social scientists have not
looked hard enough, or in the right ways, or in the right places.
It might be that the design of the ANES time series data
collections(described in more detail below) may contribute to the

And if that is so, it would be useful to consider what ANES should
do to collect data that would shed more light on the dynamics
of presidential campaigns.

With these sorts of questions in mind, the ANES Board of
Overseers, with support from the Annenberg School for
Communication of the University of Pennsylania, is convening this
fall a Conference on the Impact of the Presidential Campaign. 
The Conference will deal with two related but practically
disjoint sets of questions:  questions concerned with the
intellectual agenda that should be at the heart of this
enterprise; and questions concerned with issues of data

Issues Concerning the Intellectual Agenda

    * Why are campaigns important?  For activating political
      predispositions?  For teaching voters about candidates
      and issues?  For manipulating voters?

    * Who or what sets the campaign agenda?  Candidates?
      Voters?  Political activists?  The media?  Political
      advertisements?  Does the campaign agenda significantly
      influence the weights voters attach to specific issues,
      performance evaluations, or candidate traits in their
      final vote choices?  If so, how?

    * How important are the major predictable events of the
      campaign -- primaries, conventions, debates -- in shaping
      candidate evaluation and electoral choice?  How important
      candidate evaluation and electoral choice?  How important
      are unpredictable events  --   international crises,
      endorsements, political gaffes, and scandals?  More
      generally, what do we know about the impact of
      campaign events?  Under what circumstances is that
      impact concentrated or diffuse, transitory or long
      lasting, predictable or unpredictable?  What are the
      mechanisms by which these events affect the way citizens
      come to understand the candidates and issues?

    * How can we tell if a candidate runs a good campaign?  Is
      it possible, or necessary, to distinguish between the
      quality of a campaign and the favorability of the
      political environment in which it is conducted?

Issues Concerning Data Collection

    * The basic design of the traditional ANES Pre- /
      Post-Election Study calls for a pre-election, face-to-face
      interview with a national cross-section of citizens of
      voting-age and a follow-up interview with the same
      respondents after the election.  The pre-election
      interviewing is typically conducted in the two months prior
      to election day; the sample is comprised of four,
      self-representing quarter-samples released at two week
      intervals prior to the election.  In some Election Studies
      (1992 ANES most recently) a portion of the respondents were
      part of a panel study first interviewed two or four years

    * How adequate is this traditional ANES study design for
      understanding the nature and impact of political
      campaigns?  If it is inadequate, what sort of modification
      or supplement to the basic design should ANES consider?
      An election year panel study of the sort ANES conducted
      in 1980?  A rolling cross-section study of the sort ANES
      fielded in 1984?  An earlier start to the traditional
      pre-election interviewing?  A release of sample targeted
      to coincide with particular campaign events such as
      conventions, debates, or political gaffs?  Some combination
      of these or other features?  What is the utility of
      these alternative designs as well as those that have
      been employed by other teams of scholars who have tried
      to assess the impact of political campaigns?  What designs
      are optimal for what purposes?

    * What sorts of survey instrumentation should be developed in
      order to learn more about the impact of the campaign
      on voters?  New questions on media exposure and attention?
      More detailed questions on mobilizing activities of
      parties, interest groups and other organizations? 
      Questions measuring familiarity with prominent campaign
      themes?  Experimental manipulations attempting to impose
      alternative frames on respondents' thinking about
      candidates or issues?

    * What sorts of contextual data should be gathered in order
      to learn more about the content of the campaign?  Media
      content?  Candidate advertisements?  Travel schedules?
      Information on campaign organizations and their activities?
      Financial contributions?  Efforts at mobilization by
      social, religious, and political groups.

The Board welcomes your advice on these questions and
would appreciate suggestions about other issues that it
should consider in deliberations about the theoretical and design
issues at stake.  Two points need to be stressed here.  First,
the Board has reached no conclusions with respect to either the
appropriate intellectual agenda or the appropriate research
design for studying presidential campaigns.  The second is that
the Board does not currently have funds to support supplemental
data collection in 1996 beyond the traditional Pre- /
Post-ElectionStudy with ordered sample releases as described
above.  A central purpose of this Conference is to advise the
Board on whether it should modify the design of the 1996
Pre-/Post-Election Study data collection (in ways that would not
damage the ANES time series that dates back to 1952) and/or
whether it should seek additional funds (presumably beyond NSF)
to carry out a supplemental data collection beyond that already
funded by current NSF support.

We invite interested social scientists to send us a 2-3 page
memo or an already written paper responding to one or more
of these issues.  If a fresh paper is being proposed for
presentation at the Conference, the paper does not need to be
submitted at this time, though a brief description of its content
would be appropriate.  Proposals for what to study should
include:  (a) an argument for its theoretical and/or political
importance; (b) an argument for why the research cannot be
conducted using existing data; and (c) a clear indication of what
would have to be added to the traditional ANES Pre- /
Post-Election Study (in the way of questions, interviews, or
contextual data).

The deadline for submission of your response is August 26, 1994.
Please direct your reply to:

     Research and Development Conferences
     The National Election Studies
     Center for Political Studies/ISR
     P.O. Box 1248
     Ann Arbor, Michigan 48106-1248

You can also reply by email to: