Developing consideration set models of voting behavior
Posted 14/6/2008 17:46 (#59)
Subject: Developing consideration set models of voting behavior

New user

Posts: 1



Associate Professor
Co-Principal Investigator of the Swedish National Election Studies
Department of Political Science, University of Gothenburg, SWEDEN

Assistant Professor of Political Science
Co-Principal Investigator of the Dutch Parliamentary Election Study 2006
Department of Political Science and Research Methods
University of Twente, THE NETHERLANDS

Many well established explanatory models of party choice are waning as increasing portions of citizens in established democracies display a highly individualized and volatile voting behaviour. Decades of electoral research have shown weakening ties between parties and voters, eroding social cleavages, increasing electoral volatility, and increasing proportions of late deciders. As a result, the predictive power of existing models is slowly diminishing.

In the companion of many other scholars around the world, we seek to contribute to an ongoing effort to develop consideration set models of voting behavior to meet the challenges of an era of individualized political behavior (Oscarsson et al. 1997; Schmitt and Wessels 2005; Rosema 2006; Thomassen and Rosema 2008).

Consideration set models of electoral choice is not a novel idea. The intellectual roots originate in the well known critique against how the Michigan concept of party identification was adopted for analyses of voting behavior in European multi party contexts (Thomassen 1976). In more recent work, the consideration set models depart from existing consumer oriented models of voting behaviour, theories of information processing from political psychology and consumer choice theories from marketing research. Typically, the key assumption is that while earlier generations of citizens manifested group based interests or identifications by routinely supporting parties at elections, voters in the 21st century actually choose from a subset of available alternatives. Consequently, most citizens enter the election campaign with a consideration set or a party set that typically includes two or more alternatives that they actually will consider voting for in the election.
The full decision process is often portrayed as a two step process (analogue to consideration set approaches used in marketing research) where citizens in the first stage make an initial selection of tolerable alternatives and then, in a second stage, enter a final decision process to decide what party to vote for (Steenbergen and Hangartner 2008; de Vries and Rosema 2008). While the size and content of the consideration set is assumed to be influenced by stable long term ideological predispositions, the information processing that precede the final party choice is presumed to be heavily influenced by short term effects of factors such as campaign agendas, campaign events, debates, media coverage, and person-to-person conversation.

The current research agenda in the ongoing efforts to develop consideration set models is to invent ways to a) effectively identify the “real” set of alternatives that actually face individual voters at times of elections, b) closely monitor voters’ actual decision processes during intense election campaigns, and c) explain the outcomes of the decision processes, taking into account not only voter characteristics, but also the characteristics of the consideration set the actual decision process, and the characteristics of the alternatives included in the set. Our proposal is designed to meet the first challenge (a) of identifying what alternatives are actually present in voters’ choice process.

Many conventional survey questions have been used in the past to try isolate what parties are really included in a voters consideration set (dislike-like scales, probability to vote-scales, second best party)(Oscarsson 2004). However, none of these measures can meet the requirements of the model. It is difficult to make robust and reliable operationalisations of the concept of a party set because the size of the consideration sets either gets too large (include parties that are actually never considered) or too small (do not include a party that was actually voted for) (Oscarsson et al. 1997). Instead, many other alternative question formats are currently being evaluated by scholars in the Swedish National Election Studies Program (Oscarsson and Holmberg 2008).

The probability to vote questions, which have become a standard element in the EES, link up well to the underlying idea that electoral researchers should think beyond electoral choice as a single a categorical variable (Van der Eijk et al. 2006). However, those questions cannot tell us which parties voters did and did not consider to vote for (another way to put it: those questions do not come with a ‘cut-off point’ for parties actually considered and not considered). Hence, we need additional measures to tap that aspect of the choice process. The proposed questions do this and fit the idea that core business of studies like the EES is “measurement of the dependent variable in all its aspects” (Van der Eijk, 2002). Basically, the proposed questions add this crucial element of the choice process to the survey.

The most successful strategy seems to be to retrieve measurements during the actual decision process very close to the election; to ask voters in pre-election interviews (or in campaign panel studies) what parties they are choosing between or what parties they are actually considering at the time of the interview. Since the EES09 is a post-election telephone survey, you need to rely on recall data about what parties that were considered prior to the election. Such measures have been applied without difficulty in post-election surveys of national elections studies (e.g. Dutch Parliamentary Election Studies).

In order to identify the size (the number of parties considered) and content (the alternatives considered) of voters’ party sets, we propose to include a follow up question to the question(s) about party choice in the 2009 European Election Study. This instrument is already a part of the CSES module III. If the question is included also in the 2009 European Election Study, we will have excellent opportunities to apply the consideration set model of voting behavior both in the context of first order national elections (CSES) and the context of second order european elections (EES).

So far, previous research on consideration sets have typically included only party alternatives when defining the realistic choice options of individual voters. However, we know for certain that many voters also consider to abstain, especially in the European Parliamentary Elections. A complete consideration set approach need to include the option to abstain from voting.

Furthermore, since non-voters also may have had party alternatives in their consideration sets prior to the election, you need to ask PS1 also to respondents that report that they did not vote in the European Parliamentary Election. This way you will be able to identify the complete choice sets of all european citizens in conjunction to the European Parliamentary Election 2009.

In the proposal, we have included some standard questions (Q1 and Q2 are from the EES04) in order to get a better overview of the part of the interview that deals with the major dependents (voter turnout and party choice). Our proposal consists of three questions (Q3-Q5). However, Q4 is only asked to party voters, and Q5 is only asked to non-voters.

[A lot of people abstained in the European Parliament elections of June 12th, while others voted] Did you cast your vote?
[YES, voted]
[NO, did not vote]  Go to Q5
Don’t know
No answer

Which party did you vote for [in the European Parliamentary Election]?
[Parties]  Go to Q3
Did not Vote  Go to Q5
Do not remember
No answer

a. Did you consider voting for any other party [in the Election to European Parliament on June 12th]? [Yes/No  Go to Q6]
b. Which one?
Q3a and Q3b is repeated until the respondent says [No]

Q4 CONSIDERED NOT VOTING? (question to voters)
Did you consider NOT to vote in the European Parliamentary Election on June 12th? [Yes/No]

Q5 CONSIDERED VOTING FOR A PARTY (question to non-voters)
a. [During the Election Campaign] : Did you consider voting for [a party]/[any other party] [in the European Parliamentary Election? [Yes/No]

b. Which one?

Q4a and Q4b is repeated until the respondent says [No]


To make this proposal useful, the minimum number of questions required is the set of questions asked to party voters (Q3a and Q3b). However, it should be stressed that the length of the interview will not increase much if you choose to ask the questions to the non-voters (Q5). Asking abstainers about considered alternatives is highly relevant not only for consideration set analyses, but also for scholars that mainly focus on analyses of individual turnout.

The proposal givers have been using EES-data extensively in past publications (Oscarsson and Holmberg 2006; Rosema 2007) and since they are appointed co-prin¬cipal investigators in the national election studies program in the Nether¬lands and Sweden, they are already obliged to publish books (Oscarsson), reports and articles on the 2009 EES-study in the near future. Two manuscripts (de Vries & Rosema; Rosema & de Vries) that use EES data are in preparation for submission.

Given the interest in consideration set approach to voting behavior, it is highly likely that most scholars being cited in this proposal will contribute to the analyses of European voters’ consideration sets. Comparative EES09-data on complete choice sets for 27 European countries will not go unnoticed in the research community. This data collection will definitely advance the research on voting behavior. Furthermore, these questions may be used to further validate the probability to vote questions.

In addition, we are also preparing a panel proposal for the ECPR-meeting in Potsdam, August 2009.


de Vries, Catherine E, and Martin Rosema. 2008. "Taking voters' consideration set into consideration: Modelling electoral choice in two stages." Draft paper proposal.
Oscarsson, Henrik. 2004. "Verkligen väljande väljare [Virtually Voting Voters]." In Väljare. Svenskt väljarbeteende under 50 år [Voters. Fifty Years of Swedish Voting Behavior], ed. S. Holmberg and H. Oscarsson. Stockholm: Norstedts Juridik AB.
Oscarsson, Henrik, Mikael Gilljam, and Donald Granberg. 1997. "The Concept of Party Set. A Viable Approach or Just Another Way to Slice the Same Cheese?" Göteborg University: Department of Political Science.
Oscarsson, Henrik, and Sören Holmberg. 2006. Europaval [European Elections]. Göteborgs universitet: Statsvetenskapliga institutionen.
———. 2008. Regeringsskifte. Väljarna och valet 2006. [Government Turnover. Voting behavior in the 2006 Election]. Stockholm: Norstedts Juridik (forthcoming in october 2008).
Rosema, Martin. 2006. "Partisanship, candidate evaluations, and prospective voting." Electoral Studies 25:467-88.
———. 2007. "Low turnout: Threat to democracy or blessing in disguise? Consequences of citizens' varying tendencies to vote." Electoral Studies 26: 612-23.
Schmitt, Hermann, and Bernhard Wessels. 2005. "Under which conditions do general elections provide a meaningful choice set and what happens if they don't." In Revised Draft of a Paper for CSES III, the Third Wave of the Comparative Study of Electoral Systems
Steenbergen, M.R., and Dominik Hangartner. 2008. "Political Choice Sets in Multi-Party Elections." In Paper prepared for the conference The politics of change. Amsterdam.
Thomassen, Jacques. 1976. "Party Identification as a Cross-National Concept: Its Meaning in the Netherlands." In Party Identification and Beyond, ed. I. Budge, I. Crewe and D. Farlie. London: John Wiley & Sons.
Thomassen, Jacques, and Martin Rosema. 2008. "Party identification revisited." In Party Identification, Social Identity and Political Experience: Partisanship, ed. J. Bartle and P. Belluci. London: Routledge.
van der Eijk, Cees. 2002. "Design issues in electoral research: taking care of (core) business." Electoral Studies 21: 189-206.
van der Eijk, Cees, Wouter van der Brug, Martin Kroh, and Mark Franklin. 2006. ”Rethinking the dependent variable in voting behavior: On the measurement and analysis of electoral utilities.” Electoral Studies 25: 424-447.

Top of the page Bottom of the page
Posted 16/7/2008 17:07 (#104 - in reply to #59)
Subject: RE: Developing consideration set models of voting behavior

Posts: 26
Location: University of Oxford, UK
Dear Professors Oscarsson and Rosema,

Thank you for submitting this proposal to the Open Forum. The PIREDEU Steering Committee met at the end of June to evaluate each of the proposals. We assessed them on the basis of whether they met the following criteria:

* An explicit argument about why the proposed question/coding category merited inclusion in one or more of the PIREDEU data components.

* An explicit argument about the conceptual and theoretical foundations of the question/coding category.

* An explicit case for how the question/coding category facilitates integration and linking of several data components. The PIREDEU Steering Committee preferred proposals that allowed for conceptual integration across the five data components (i.e. voter survey, candidate survey, media study, manifestos and contextual data).

* An explicit consideration of how the proposed question/coding category linked with questions/coding categories in past data collection efforts.

The PIREDEU Steering Committee favoured proposals that ensured over time and across instrument comparability. Moreover, given that the voter and candidate surveys can only contain a limited number of question items, priority was given to proposals with succinct question formats.

On this basis we ranked each question in the proposal as follows:

(1) The proposed item will be included in data collection instrument
(2) High priority proposal that will be included if space and time constraints permit
(3) Proposal can only be included if additional funding is secured
(4) Proposed item is not a priority

The item(s) from your proposal received the following ranking(s):

Instrument: Voter Survey

Q1 VOTED [A lot of people abstained in the European Parliament elections of June 12th, while others voted] Did you cast your vote?
Ranking: 4
Q2 PARTY CHOICE Which party did you vote for [in the European Parliamentary Election]?
Ranking: 4
Q3 CONSIDERED VOTING FOR A PARTY? a. Did you consider voting for any other party [in the Election to European Parliament on June 12th]? Go to Q6] [Yes/No] b. Which one? Q3a and Q3b is repeated until the respondent says [No]
Ranking: 4
Q4 CONSIDERED NOT VOTING? (question to voters) Did you consider NOT to vote in the European Parliamentary Election on June 12th?
Ranking: 4
Q5 CONSIDERED VOTING FOR A PARTY (question to non-voters) a. [During the Election Campaign] : Did you consider voting for [a party]/[any other party] b. Which one? Q4a and Q4b is repeated until the respondent says [No]
Ranking: 4

Rationale: The propensity to vote questions are already included in the core battery because they are part of previous EES studies. The steering committee decided to include a question for non-voters, but the final wording will be different than Q4.

Thank you again for your participation in this process. We hope that you will continue to use the Open Forum to comment on the questionnaires/codebook that will be posted online on the Forum in the autumn.

Yours sincerely,

Mark Franklin
Chair of the PIREDEU Steering Committee

Sara Hobolt
Deputy Chair of the PIREDEU Steering Committee

Top of the page Bottom of the page