Professional Documents
Culture Documents
1, 2014
Voting Advice Applications (VAAs) have experienced a rising demand in recent years, being an
efficient and reliable way for voters to gain political information prior to elections. This article uses
an online survey to examine the impact of the German VAA (the “Wahl-O-Mat”) on the political
knowledge of its users concerning party positions during the 2009 German Federal Election. The
article first assesses empirically the extent to which Wahl-O-Mat users are able to correctly identify
party positions on relevant issues, and the extent to which they are misinformed. Multivariate
analyses with a path model for categorical data are then undertaken to establish whether Wahl-O-
Mat usage enhances political knowledge about party positions. It is shown that Wahl-O-Mat use
has a positive effect on political knowledge, indicating that electoral effects can be partially explained
by this change in the information level of users. However, it is also shown that political knowledge
is very heterogeneous, and the level of confusion about party positions is high.
KEY WORDS: Voting Advice Applications, VAAs, Wahl-O-Mat, political knowledge, party positions,
misinformation, 2009 German Federal Election, political online communication
Introduction
46
1944-2866 # 2014 Policy Studies Organization
Published by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington Road, Oxford, OX42 DQ.
Schultze: Effects of VAAs on Political Knowledge About Party Positions 47
as Voting Advice Applications (VAAs), have the explicit goal to inform citizens
about the relevant policy positions of political parties and to motivate them to
political participation prior to elections. These applications share a basic
functionality: VAAs compare the voters’ policy positions with those of parties or
candidates running for election (Cedroni & Garzia, 2010; Garzia &
Marschall, 2012). After voters have marked their positions on a list of policy
statements, VAAs compare their answer patterns with the patterns of the parties/
candidates, indicating which party or candidate has the highest degree of
proximity to the users’ positions.
Because of their focus on political topics such tools clearly relate to the issue-
orientation of citizens: The users of these tools are not only made aware of the
relevant topics of the election campaign, but they also catch up on the parties’
positions in a comparative perspective, which should lead to an increase in the
political knowledge of VAA users in terms of party positions. The German VAA,
the Wahl-O-Mat, is produced by the Federal Agency for Civic Education
(“Bundeszentrale für politische Bildung”), a governmental agency connected to
the Ministry of the Interior. A first Wahl-O-Mat version was provided for the
2002 Federal Election, followed by versions for the Federal Elections in 2005, 2009,
and 2013, and the European Elections in 2004, 2009, and 2014. Already in its first
run in 2002, the tool was used 3.6 million times. Before the Federal Election in
2009, the reference point of this study, the tool provided 6.7 million advices
(Marschall, 2011, p. 137), presumably reaching a remarkable share of the German
electorate.
This article analyzes empirically whether the German VAA has the expected
effect on political knowledge about party positions, based on survey data for the
2009 German Federal Election. The following section covers the theoretical
framework and the empirical results for political knowledge in Germany, as well
as results on the impact of VAAs on political behavior in the context of political
online communication. After a brief description of the data, methods and
operationalization of political knowledge, descriptive analyses show the level of
knowledge and misinformation about party positions. Furthermore, multivariate
methods reveal whether the usage of the Wahl-O-Mat has a positive effect on the
political knowledge. In the final section, the findings are summarized and
discussed within the context of “media effects” in political online communication
(Lin, 2009) and for political education.
There are only a few findings based on survey research for the level of
political knowledge, its determinants, and consequences in Germany. In contrast
to the United States and the United Kingdom, where this topic has been covered
more extensively within a lasting tradition (Andersen, Tilley, & Heath, 2005; Delli
Carpini & Keeter, 1996; Luskin & Bullock, 2011), German social scientists have
48 Policy & Internet, 6:1
For example, Westle (2005, p. 507) estimates the share of voters that have chosen
their party based on misinformation about policy positions at about 20 percent in
the Western and 30 percent in the Eastern part of Germany for the 2002 Federal
Election. According to these findings, it is important not only to take the level of
political information into account, but also the level of misinformation among
the electorate.
Therefore, a differentiation between correct answers (“objective knowl-
edge”) and incorrect answers (“subjective knowledge”) is helpful (Vetter &
Maier, 2005, p. 55; Westle, 2005), supplemented by the answer category “do not
know” explicitly expressing a lack of knowledge. For the individual voting
decision, a low objective political knowledge about party positions is not
necessarily problematic, because such a decision can be based on numerous
heuristics like attitudes toward the political candidates and their attributes, or a
social group tie or party attachment, which are all important determinants for
explaining voting behavior in Germany (Arzheimer & Schoen, 2007; Keller-
mann & Rattinger, 2005; Roßteutscher, 2012; Schoen, 2011). Even for issue-
orientated voters, broad knowledge is neither necessary nor realistic. Already
in 1964, Converse (1964, p. 245) showed that the electorate is divided into
various issue publics that are only interested in certain policy areas and that
possess a kind of expert knowledge only in their preferred policy fields (Anand
& Krosnick, 2003). In this context, Krosnick (1990, p. 82) refers to the people as
“cognitive misers” when pointing out that voters based their voting decisions
only on a few salient topics. This illustrates that even issue-orientated voters do
not necessarily need a broad political knowledge for a reasonable vote choice,
as long as they are aware of their lack of knowledge instead of being
misinformed.
In this regard, the mass media play a central role, because they create the
public sphere in which the voters can seek information and discuss political
topics and positions (Neidhardt, 1994). With the diffusion of the Internet, the
classical media system has changed and expanded remarkably (Gurevitch,
Coleman, & Blumler, 2009, p. 169; Schulz, 2011, p. 19). The range of political
information available for citizens online has never been higher than today,
although its share is small compared to the enormous amount of nonpolitical
content on the Web (Emmer, 2005; Emmer & Wolling, 2010).
Indeed, there are different empirical findings concerning whether the Internet
is used as a substitute (van Eimeren & Frees, 2011, p. 343; Faas & Partheymüller,
2011) or is complementary to other media (Emmer, Vowe, & Wolling, 2011;
Kolo, 2010) for political information and communication. However, it is not in
doubt that these new media cause changes in individual communication
behavior. Investigating the “media effects” (Lin, 2009) of Internet applications is a
rather challenging endeavor due to the heterogeneous formats of such applica-
tions. Because the Internet is a technical operating system and a “first-order
medium,” encompassing numerous “second-order media” characterized by
semiotic, institutional, and organizational heterogeneity (Beck, 2010, p. 17),
Schultze: Effects of VAAs on Political Knowledge About Party Positions 51
potential effects can vary in their nature and strength and should therefore be
investigated specifically for each application.
The German VAA, as a form of political online communication (Emmer
et al., 2011), is such a second-order medium of the Internet that is analyzed here
using such an application-focused approach. The Wahl-O-Mat and other VAAs
have the potential to increase the political information level of the electorate and
to reduce the level of misinformation by making people aware of relevant issues
prior to elections.
For this purpose, the German VAA offers a selection of 38 policy issues that
statistically differentiate the political parties running for election. The parties are
asked to position themselves on a long list of policy issues and to provide reasons
for their positions. The policy issues are in the form of short statements that are
worked out by an editorial board consisting of young and first-time voters, social
scientists, and experts from various disciplines. Party manifestos and party
statements serve as the basis for the selection and formulation of the statements
(Marschall, 2011). The statements are written in a short and clear way, so that
users can easily understand them and take their positions (e.g., “genetically
manipulated food should not be produced in Germany”). For each issue, the user
can choose to agree or disagree with the statement, to take a “neutral” stance, or
to skip it. It is possible to go back to an issue and change the stance at any time.
Before the tool calculates the degree of proximity between the users’ positions
and the parties’ positions, the users can choose to weight the statements they
consider more important. These count double in the proximity calculation. After
this step, users can choose up to eight parties out of a list of all parties registered
for the election and compare the results. For the parties chosen, the tool calculates
the proximity between the self-positioning of the user and each party according
to the city-block method (Marschall & Schmidt, 2010). The result is presented as a
ranking list, beginning with the party with the highest proximity to the user’s
position. Moreover, the positions of all the parties chosen for the comparison are
listed in a form that allows the users to compare their positions (see the
Appendix). The users can also choose to read the reasoning of the parties. The
potential effects of VAAs arise due to this special format: They confront citizens
with numerous issues about which the users might otherwise not have been
informed or explicitly positioned themselves. As a consequence, the tools may to
a certain degree be able to resolve the selectivity of political information seeking
in the Internet, that otherwise proceeds along the issue publics of the voters
(Kim, 2009).
Whether such tools can meet the expectation of informing people in a broader
sense is rarely covered by research from an international perspective. Early
research about these tools has been focused on more descriptive questions
concerning characterization of users, and how they differ in terms of their socio-
demographic backgrounds and political attitudes from the entire electorate, but
also compared to the online population (Hooghe & Teepe, 2007; Marschall, 2011;
Wall, Sudulich, Costello, & Leon, 2009). In an international perspective, the users
of the tools can be consistently characterized as young, well educated, and
52 Policy & Internet, 6:1
politically interested, with a slightly higher share of men using the tool (De
Rosa, 2010; Marschall & Schultze, 2012c; Wall et al., 2009). With the growing
popularity of VAAs—evidenced by the rising number of these tools across
Europe and the increasing share of voters that use them—the attention paid to
the quality of results they generate has increased. Therefore, another area of VAA
research has focused on the impact of different calculation methods on the
presented results (Louwerse & Rosema, 2013; Mendez, 2012), as well as on
methodological questions about the assessment of policy statements and different
possibilities to map party positions with such information (Ladner, Felder,
Gerber, & Fivaz, 2010; Ramonaitė, 2010; Walgrave, Nuytemans, & Pepermans,
2009).
However, the majority of studies have focused on potential VAA effects on
the political behavior of users. Taking the German case as an example, it has
been shown that the Wahl-O-Mat had an effect on voter turnout for the 2009
German Federal Election: Having controlled specific characteristics of the users
of this tool, as well as other variables that usually explain the intention to vote,
the usage of the Wahl-O-Mat had a significant positive effect on the probability
of going to the ballot box (Marschall & Schultze, 2012a). This result supports
other studies gauging the mobilization capacity of VAAs (Ladner &
Pianzola, 2010; Mykkänen & Moring, 2006). Beside the influence on the
intention to vote, a number of studies have analyzed a possible change of the
voting decision due to the use of these Internet applications, reaching
ambiguous conclusions concerning the strength of such effects (Ladner &
Pianzola, 2010; Marschall, 2005; Mykkänen & Moring, 2006; Wall, Krouwel, &
Vitiello, 2012).
In contrast to this large body of studies dealing with the impact of these tools
on political behavior, the possible effects of VAAs on political knowledge have
been seldom studied, or at best addressed tangentially (Marschall &
Schmidt, 2010; Pianzola, 2011). With the exception of a study about the Swiss
VAA “smartvote” (Fivaz & Nadig, 2010), this effect has so far—to my best
knowledge—not been taken into account systematically as a central research
question. For the German case, this study is therefore of an explorative nature.
The expectation regarding the relationship of Wahl-O-Mat usage and political
knowledge is rather obvious: The Wahl-O-Mat can increase political knowledge
about party positions by confronting users with numerous issues and a
comparison of their own stances and the policy positions of the parties. This
should lead to a learning effect among the users about political topics, especially
beyond their individual issue publics. In this respect, the process of knowledge
acquisition via VAAs depends on the form of political knowledge. Specific
knowledge about party positions should be primarily gained by comparing and
memorizing the policy positions on the result screen of the Wahl-O-Mat, while
general political knowledge should be generated by discussing the results with
other people, especially a surprising party sequence in the proximity calculations
(Marschall & Schmidt, 2010). A possible learning effect through collecting and
processing new political information is thereby an important precondition for an
Schultze: Effects of VAAs on Political Knowledge About Party Positions 53
explanation of a change in the voting intention and voting decision, and a step
forward in terms of identifying and understanding cognitive processes in voting
behavior. Furthermore, it can be shown (at least for the Wahl-O-Mat), whether
and under which conditions political online communication can have an impact
on citizens.
analyses. These three political issues were relevant prior to the 2009 German
Federal Election, and the parties of analysis (CDU, SPD, FDP, Greens, and Left
Party) had clear positions on these policy proposals in the election campaign.
Statements about these topics were also included in the Wahl-O-Mat using the
identical wording used in the online survey and were answered by the parties
according to their election manifestos. However, it must be acknowledged that
these knowledge questions about the three issues in the GLES online survey do
not cover the full range of statements that are presented in the Wahl-O-Mat, but
are the only variables that are included in the GLES data set for studying political
knowledge about party positions. The use of the Wahl-O-Mat, including the
observation of the parties’ positions toward these issues on the results page,
should have a corresponding positive effect on the political knowledge of German
VAA users about these positions.
In the GLES online survey, respondents were asked these political knowl-
edge questions as closed questions with the response categories if the parties
“agree,” “disagree,” or have “neutral” positions on these statements.5 Further-
more, it was possible to express a lack of knowledge with the “do not know”
category. Due to these answer possibilities, it can be differentiated between
correct answers, incorrect answers, and a reflexive lack of knowledge (i.e.,
explicit “do not know”) of the respondents (Vetter & Maier, 2005, p. 55;
Westle, 2005). Table 1 summarizes the exact wording of the statements that are
used in the Wahl-O-Mat as well as in the GLES survey and shows the party
positions on these issues.
Table 1. Wording of the Statements in the Wahl-O-Mat, Respectively, in the GLES Survey and Parties’
Positions
Notes: Wording in the GLES survey is identical with the wording in the Wahl-O-Mat. The
following question was asked in the GLES survey initially, before asking about the
perceived positions of the parties: “Bitte geben Sie zu den einzelnen Vorschlägen jeweils
an, ob die einzelnen Parteien diese Ihrer Meinung nach eher unterstützen, eher ablehnen
oder ob sie ihnen neutral gegenüber stehen?” [“Please indicate to each of the proposals, if
in your opinion the parties rather support, rather decline or have a neutral position?”].
Schultze: Effects of VAAs on Political Knowledge About Party Positions 55
Descriptive results about the distribution of correct and incorrect answers are
presented next, for the three policy positions of the five parties. The correct
answers are then summated, categorized, and related to the use of the Wahl-O-
Mat.
Table 2 shows the share of correct and incorrect answers of the respondents
for the 15 knowledge questions. Most widely known, with 71 percent correct
answers, is the Greens’ position against a lifespan extension for nuclear power
plants, which can be explained by the linkage of this policy position with the
party’s creation, and by the issue ownership of the Greens concerning environ-
mental issues in the German electorate. Most unknown is the position of the FDP
to this policy statement: Only 37.3 percent of the respondents know that the
liberal party supports a lifespan extension of such plants, which they extensively
communicated in the 2009 election campaign.
While 61.2 percent of the respondents can judge the position of the CDU for a
lifespan extension correctly, the position against such an extension of the SPD is
only known by 53.6 percent of the German online population. Better known is the
position of the Social Democrats for the introduction of a minimum wage, a key
issue of their 2009 election campaign. In general, the share of correct answers for
this issue is similar compared to the question about nuclear power plants, but
shows a different distribution across the parties. A comparable number of
respondents can correctly identify the position of the SPD (60.9 percent) and the
Left Party (58.9 percent) in support of the introduction of a minimum wage. On
the other hand, with about 40–50 percent correct answers, the level of knowledge
for the other parties (CDU, FDP, and Greens) is considerable lower. In contrast to
these two issues, the policy positions of any parties concerning a prohibition of
the use of governmental spyware are widely unknown in the online electorate, as
the very low level of information indicates. Lowest in this respect are the correct
evaluations of the FDP and SPD toward this topic: Only 19.8 percent knew that
Table 2. Correct and Incorrect Answers About Policy Positions of the Parties in Percent
the FDP supported and only 20.2 percent knew that the SPD was against a
prohibition of such spyware in the context of the 2009 German Federal Election.
All in all, the share of correct answers ranges between 20 and 70 percent for
the knowledge questions. However, it must be acknowledged that the respon-
dents represent the online electorate during the final stage of the electoral
campaign. This group has, compared to the entire electorate, a higher educational
attainment on average, and a higher interest in politics (Marschall &
Schultze, 2012b), which normally should lead to a higher level of political
information. Therefore, it can be assumed that a representative survey of the
entire German voting population would reveal an even lower level of political
knowledge.
Table 2 also shows the share of respondents who are misinformed about
party positions, reflected by incorrect answers. The frontrunner is the position of
the SPD against a prohibition of the use of governmental spyware: 38.5 percent
are misinformed about this position, clearly outnumbering persons who are
correctly informed (20.2 percent). A similar discrepancy between correct and
incorrect answers can be found for positions of the FDP: The share of
misinformed respondents is relatively high for all policy statements (between 27.3
and 36.0 percent). Concerning the question of a prohibition of the use of
governmental spyware, the share of misinformed people outnumbers the share of
respondents who correctly answered (36 vs. 19.8 percent).
The smallest amount of misinformation can be observed for the Greens’
position against nuclear power plants (9 percent), and all positions of the Left
Party (between 11.1 and 21.2 percent). In general, the range of misinformation
varies enormously between 9 and 38.5 percent. The “do not know” category was
extensively used, ranging from 20 to 45.5 percent. Taking into account that these
issues were intensively covered in the mass media, the amount of misinformation
and do not knows immediately before the 2009 Federal Election is high among the
online electorate.
For the further analyses, only the correct answers are taken into account.
Figure 1 shows the degree of political knowledge about party positions by
summating the number of correct answers of the 15 knowledge questions for all
respondents. In the mode category, “0 correct answers” are mainly people
represented who either used the “do not know” category extensively or who
refused to answer this battery of questions. It is controversial how to deal with
such respondents. To avoid a too-optimistic distribution of the level of political
information, these cases were not dropped from the analyses, but were not
merged either with other categories in this study. Altogether, 16.1 percent of the
sample population can be classified in this category.
Following the methodological approach of Roy (2009, p. 125), the residual
respondents were arranged in three groups to represent the differences in their
levels of political information. According to this categorization, 23.9 percent of the
respondents have a low political knowledge about party positions (1–5 correct
answers), 33.5 percent possess a medium political knowledge (6–10 correct
answers), and 26.5 percent have a high level of information about party positions
Schultze: Effects of VAAs on Political Knowledge About Party Positions 57
(11–15 correct answers). Therefore, the political knowledge of the online popula-
tion can be considered as exceedingly heterogeneous shortly before the 2009
Federal Election.6 Reasons for this heterogeneity may lie in the individual
belonging to one or more issue publics that are covered with these knowledge
questions, and that would favor political knowledge in these policy areas while
ignoring other policy fields (Krosnick, 1990). Furthermore, different sociodemo-
graphic characteristics such as political interest and education, as well as the
quality of sources used for political information, influence political knowledge
and explain the heterogeneity among the electorate (Vetter & Maier, 2005,
pp. 72–3). In the next section, the Wahl-O-Mat is integrated in the analyses. About
38.5 percent of the respondents reported having used the Wahl-O-Mat for the
2009 German Federal Election, demonstrating the popularity of the German VAA.
Table 3 shows the groups with different levels of information about party
positions cross tabbed with the use of the Internet application. The bivariate
frequency table draws a clear picture: Wahl-O-Mat users are more likely to have a
high political knowledge about party positions. From those respondents who
have used the tool, about 39.3 percent belong to the group of the highly informed
Table 3. Cross Tab for Political Knowledge About Party Positions and Wahl-O-Mat Usage
(Column Percent)
people, and only 18.9 percent have a low political knowledge; only 7 percent of
Wahl-O-Mat users answered all questions incorrectly or refused to answer at all.
For the nonusers of the Wahl-O-Mat, the latter share, at 21.1 percent, is about
three times as high compared with users. Nonusers are disproportionally highly
represented in the group of people with low political information, while the
proportion of users and nonusers with a medium knowledge about party
positions is similar. Only 18.7 percent of nonusers can be characterized as being
highly politically informed in this context.
To summarize, the bivariate analysis shows that Wahl-O-Mat users have a
higher political knowledge about party positions than nonusers. The question of
whether this relationship is robust, which in this context means persistent when
controlled for relevant third variables, needs to be analyzed with more elaborate
multivariate analyses. These are presented in the next section.
Figure 2. Path Model for the Explanation of Political Knowledge About Party Positions.
Notes: p < 0.01. The figure shows the coefficients in the following way: unstandardized coefficient
(standard error) standardized coefficient. Only standardized coefficients are reported in the text.
For more complex models taking recursive effects into account and for
disentangling the causal relationships between those variables, studies that use
panel data or experimental designs (Pianzola, Trechsel, Schwerdt, Vassil, &
Alvarez, 2012) are needed. Acknowledging these assumptions and limitations, the
presentation of results of the nonrecursive path model in Figure 2 follows the
recommendation of Kline (2011, p. 40): Unstandardized coefficient (standard
error) standardized coefficient. The standardized estimates can be compared
directly for an analysis of the strength of the effects.
First of all it can be observed that all path coefficients—with the exception of
gender to Wahl-O-Mat usage—are significant at the 1 percent level and that the
including variables in the model provide a good explanatory power for the
political knowledge about party positions, as indicated by the pseudo-R2
(McKelvey and Zavoina) of about 0.4. In the model, age shows a very interesting
effect: It exerts a moderate direct positive effect on political knowledge, as well as
an indirect positive effect via political interest. Furthermore, a higher age has a
negative effect on the usage of the German VAA. On the contrary, the usage of
the Wahl-O-Mat itself has a moderate positive effect on political knowledge,
similar to the effect of political interest on political knowledge.
Formal educational attainment has, compared to the previous effects, a lower
direct impact on political knowledge, but mainly shapes the level of information
about party positions via the indirect path of political interest. However, the
influence of education on the usage of the German VAA is comparatively low.
Major determinants of the Wahl-O-Mat usage are therefore rather a high political
interest and a younger age, and rather not educational attainment. Furthermore,
gender has no impact on the use of the tool in this multivariate perspective, but it
does on political knowledge.
Male respondents have, compared to their female counterparts, a higher
interest in politics and a higher political knowledge about party positions. Also
taking into account the moderate indirect effect from gender via political interest
to political knowledge, this leads to the conclusion that gender has a strong total
impact on political knowledge, as shown in Table 4. Education and especially
political interest also have a strong total effect on political knowledge of party
positions, while the overall effect of age is comparatively low. This is due to the
negative path from age via Wahl-O-Mat to political knowledge, weakening the
overall effect of age, despite the other positive direct and indirect paths from age
to political knowledge.
To summarize: It can be observed that the Wahl-O-Mat has a moderate
positive effect on political knowledge, controlling for important variables in a
multivariate perspective, which empirically supports the formulated expectation
concerning a learning effect of the tool about party positions.
Notes: p < 0.01. The coefficients are presented in the following way: unstandardized
coefficient (standard error) standardized coefficient.
making at the ballot boxes. Thus, low informed citizens take into account fewer
and simpler considerations for casting a vote, while highly informed voters rely
on more complex considerations for their voting decision.
Findings of political knowledge also have important implications for political
education. It is controversial what level of knowledge about political facts is
necessary or desirable, and how to integrate and measure more complex forms of
political knowledge in order to obtain a realistic picture of the political
sophistication of citizens. However, a high level of misinformation in the
electorate, leading to decisions against their own political preferences, is not
compatible with the claim of an enlightened citizenry in Western democracies.
Therefore, not only political actors and the mass media are responsible to make
politics and political processes understandable and transparent, but also citizens
should inform themselves better about politics (Trankovits, 2012). The Wahl-O-
Mat and similar tools throughout Europe are in this respect an important means
by which citizens can become informed, at least about party positions, in an
efficient, time-saving, and neutral way. The enormous public demand of VAAs,
reaching a remarkable share of the electorate in many European countries
(Cedroni & Garzia, 2010), underscores the value of such tools for political
education and participation. By confronting the citizenry with relevant policy
statements, such tools contribute toward a more enlightened and more issue-
orientated electorate.
Notes
The author would like to thank the three anonymous reviewers and the editors of Policy & Internet for
valued comments on a previous draft of this article, as well as Susan Djahangard-Mahboob, Melissa
Schiefer, and Nadine Zwiener for their support in preparing the manuscript.
1. http://gles.eu/wordpress/english/.
2. http://www.respondi.com/en/home.html.
3. Information on the sampling and weighting of the sample can be found in the method report:
www.gesis.org/wahlen/gles/daten-und-dokumente/daten/. The data set used for the analyses in
this article can also be downloaded from this website.
4. The report can be found online: www.initiatived21.de/wp-content/uploads/alt/08_NOA/
NONLINER2008.pdf.
5. All party positions were clearly for or against the formulated policy statements, therefore answers
of respondents stated a “neutral position” were treated as incorrect answers.
6. It must be mentioned that with this common procedure of using a summation index, issue- and
party-related variations in the political knowledge of the respondents could not been taken into
account in the further analyses.
7. The variable that measures the political knowledge in these analyses is the grouped number of
correct answers to the 15 knowledge questions, differentiating between varying levels of
information of the respondents. The decision to use the grouped number of correct answers and
therefore a categorical variable instead of the ungrouped number of correct answers as a
continuous variable has two reasons: First, to avoid a model with a combination from linear and
64 Policy & Internet, 6:1
probit regression coefficients that would be estimated treating the dependent variable as
continuous and second to avoid biased estimates for this continuous dependent variable in such a
model, because the distribution of the ungrouped variable that count the number of correct
answers is far away from normal.
8. Age is integrated in these analyses by dividing it by 10 to avoid an extreme range of values for
this variable.
9. Reference category for gender is “female.”
References
Anand, S., and J.A. Krosnick. 2003. “The Impact of Attitudes Toward Foreign Policy Goals on Public
Preferences Among Presidential Candidates: A Study of Issue Publics and the Attentive Public in
the 2000 U.S. Presidential Election.” Presidential Studies Quarterly 33 (1): 31–71.
Andersen, R., J. Tilley, and A.F. Heath. 2005. “Political Knowledge and Enlightened Preferences: Party
Choice Through the Electoral Cycle.” British Journal of Political Science 35 (2): 285–302.
Arzheimer, K., and H. Schoen. 2007. “Mehr als eine Erinnerung an das 19. Jahrhundert? Das
sozioökonomische und das religiös-konfessionelle Cleavage und Wahlverhalten 1994–2005.” In
Der gesamtdeutsche Wähler, ed. H. Rattinger. Baden-Baden: Nomos, 89–112.
Beck, K. 2010. “Soziologie der Online-Kommunikation.” In Handbuch Online-Kommunikation, eds. W,
Schweiger and K. Beck. Wiesbaden: VS Verlag, 15–35.
Cedroni, L., and Garzia, D. eds. 2010. Voting Advice Applications in Europe: The State of the Art. Napoli:
Scripta Web.
Converse, P.E. 1964. “The Nature of Belief Systems in Mass Publics.” In Ideology and Discontent,
ed. D.E. Apter. New York: The Free Press of Glencoe, 206–61.
Converse, P.E. 2000. “Assessing the Capacity of Mass Electorates.” Annual Reviews of Political Science 3:
331–53.
Dalton, R.J. 2008. Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies,
5th ed. Washington, DC: CQ Press.
Debus, M. 2007. “Bestimmungsfaktoren des Wahlverhaltens in Deutschland bei den Bundestagswahlen
1987, 1998 und 2002: Eine Anwendung des Modells von Adams, Merrill und Grofman.” Politische
Vierteljahresschrift 48 (2): 269–92.
Delli Carpini, M.X., and S. Keeter. 1996. What Americans Know About Politics and Why It Matters. New
Haven: Yale University Press.
De Rosa, R. 2010. “cabina-elettorale.it (Provides Advice to Italian Voters Since 2009).” In Voting Advice
Applications in Europe: The State of the Art, eds. L. Cedroni and D. Garzia. Napoli: Scripta Web,
187–212.
Downs, A. 1957. An Economic Theory of Democracy. New York: Harper.
Emmer, M. 2005. Politische Mobilisierung durch das Internet? Eine kommunikationswissenschaftliche
Untersuchung zur Wirkung eines neuen Mediums. München: Fischer Verlag.
Emmer, M., G. Vowe, and J. Wolling. 2011. Bürger Online: Die Entwicklung der politischen Online-
Kommunikation in Deutschland. Konstanz: UvK.
Emmer, M., and J. Wolling. 2010. “Online-Kommunikation und politische Öffentlichkeit.” In Handbuch
Online-Kommunikation, eds. W. Schweiger and K. Beck. Wiesbaden: VS Verlag, 36–58.
Faas, T., and J. Partheymüller. 2011. “Aber jetzt?! Politische Internetnutzung in den Bundestagswahl-
kämpfen 2005 und 2009.” In Das Internet im Wahlkampf: Analysen zur Bundestagswahl 2009, eds. E.J.
Schweitzer and S. Albrecht, Wiesbaden: VS Verlag, 119–35.
Fivaz, J., and G. Nadig. 2010. “Impact of Voting Advice Applications (VAAs) on Voter Turnout and
Their Potential Use for Civic Education.” Policy & Internet 2 (4): 167–200.
Fournier, P., A. Blais, R. Nadeau, E. Gidengil, and N. Nevitte. 2003. “Issue Importance and
Performance Voting.” Political Behavior 25 (1): 51–67.
Garzia, D., and S. Marschall. 2012. “Voting Advice Applications Under Review: The State of Research.”
International Journal of Electronic Governance 5 (3–4): 203–22.
Schultze: Effects of VAAs on Political Knowledge About Party Positions 65
Gibson, R.K., L. Wainer, and S. Ward. 2005. “Online Participation in the UK: Testing a ‘Contextualised’
Model of Internet Effects.” British Journal of Politics and International Relations 7 (4): 561–83.
Gurevitch, M., S. Coleman, and J.G. Blumler. 2009. “Political Communication: Old and New Media
Relationships.” The ANNALS of the American Academy of Political and Social Science 625: 164–81.
Hayes, A.F., M.D. Slater, and L.B. Snyder. 2008. The SAGE Sourcebook of Advanced Data Analysis Methods
for Communication Research. Los Angeles: Sage Pubn Inc.
Herrmann, M. 2008. “Moderat bevorzugt, extrem gewählt: Zum Zusammenhang von Präferenz
und Wahlentscheidung in räumlichen Modellen sachfragenorientierten Wählens.” Politische
Vierteljahresschrift 49 (1): 20–45.
Hirzalla, F., L. van Zoonen, and J. de Ridder. 2011. “Internet Use and Political Participation: Reflections
on the Mobilization/Normalization Controversy.” The Information Society 27: 1–15.
Hooghe, M., and W. Teepe. 2007. “Party Profiles on the Web: An Analysis of the Log Files of
Nonpartisan Interactive Political Internet Sites in the 2003 and 2004 Election Campaigns in
Belgium.” New Media and Society 9 (6): 965–85.
Johann, D. 2012. “Direkte und indirekte Effekte spezifischen politischen Wissens auf politische
Partizipation.” Politische Psychologie 2 (1): 23–46.
Karp, J.A. 2006. “Political Knowledge and Electoral Rules: Comparing Mixed Member Proportional
Systems in Germany and New Zealand.” Electoral Studies 25: 714–30.
Kellermann, C., and H. Rattinger. 2005. “‘Round Up the Usual Suspects’: Die Bedeutung klassischer
Bestimmungsfaktoren der Wahlentscheidung bei den Bundestagswahlen 1994 bis 2002.” In
Wahlen und Wähler, ed. J.W. Falter. Wiesbaden: VS Verlag, 189–212.
Kim, Y.M. 2009. “Issue Publics in the New Information Environment: Selectivity, Domain, Specificity,
and Extremity.” Communication Research 36 (2): 254–84.
Kline, R.B. 2011. Principles and Practice of Structural Equation Modeling, 3rd ed. New York: Guilford
Press.
Kolo, C. 2010. “Online-Medien und Wandel: Konvergenz, Diffusion, Substitution.” In Handbuch Online-
Kommunikation, eds. W. Schweiger and K. Beck. Wiesbaden: VS Verlag, 283–307.
Kraft, P., and R. Schmitt-Beck. 2013. “Helfen politische Gespräche, ‘korrekt’ zu wählen? Eine Analyse
zur Bundestagswahl 2009.” In Zivile Bürgergesellschaft und Demokratie: Aktuelle Ergebnisse der
empirischen Politikforschung, eds. S.I. Keil and S.I. Thaidigsmann. Wiesbaden: Springer, 117–38.
Krosnick, J.A. 1990. “Government Policy and Citizen Passion: A Study of Issue Publics in
Contemporary America.” Political Behavior 12 (1): 59–92.
Ladner, A., G. Felder, S. Gerber, and J. Fivaz. 2010. Die politische Positionierung der europäischen Parteien
im Vergleich: Eine Analyse der politischen Positionen der europäischen Parteien anlässlich der Wahlen des
Europäischen Parlaments 2009 mit besonderer Berücksichtigung der Schweizer Parteien. Cahier de
l’IDHEAP 252, Chavannes-Lausanne, Institut de hautes études en administration publique.
Ladner, A., and J. Pianzola. 2010. “Do Voting Advice Applications Have an Effect on Electoral
Participation and Voter Turnout? Evidence From the 2007 Swiss Federal Elections.” In Electronic
Participation, Vol. 1, eds. E. Tambouris, A. Macintosh, and O. Glassey. Berlin/Heidelberg/New
York: Springer, 211–24.
Lau, R.R., D.J. Andersen, and D.P. Redlawsk. 2008. “An Exploration of Correct Voting in Recent U.S.
Presidential Elections.” American Journal of Political Science 52 (2): 395–411.
Lau, R.R., and D.P. Redlawsk. 1997. “Voting Correctly.” American Political Science Review 91 (3): 585–98.
Lin, C.A. 2009. “Effects of the Internet.” In Media Effects: Advances in Theory and Research, eds. J. Bryant
and M.B. Oliver. New York: Routledge, 567–91.
Louwerse, T., and M. Rosema. 2013. “The Design Effects of Voting Advice Applications: Comparing
Methods of Calculating Matches.” Acta Politica. doi: 10.1057/ap.2013.30.
Luskin, R.C. 2002. “From Denial to Extenuation (and Finally Beyond): Political Sophistication and
Citizen Performance.” In Thinking About Political Psychology, eds. J.H. Kuklinksi. Cambridge:
Cambridge University Press, 281–305.
66 Policy & Internet, 6:1
Luskin, R.C., and J.G. Bullock. 2011. “‘Don’t Know’ Means ‘Don’t Know’: DK Responses and the
Public’s Level of Political Knowledge.” The Journal of Politics 73 (2): 547–57.
Maier, J. 2009. “Was die Bürger über Politik (nicht) wissen—und was die Massenmedien damit zu tun
haben: Ein Forschungsüberblick.” In Politik in der Mediendemokratie, eds. F. Marcinkowski and B.
Pfetsch. Wiesbaden: VS Verlag, 393–414.
Maier, J., A. Glantz, and S. Bathelt. 2009. “Was wissen die Bürger über Politik? Zur Erforschung der
politischen Kenntnisse in der Bundesrepublik Deutschland 1949 bis 2008.” Zeitschrift für
Parlamentsfragen 40 (3): 561–79.
Marschall, S. 2005. “Idee und Wirkung des Wahl-O-Mat.” Aus Politik und Zeitgeschichte B 51–52: 41–6.
Marschall, S. 2011. “Nutzer und Nutzen: Der Wahl-O-Mat zur Bundestagswahl 2009.” In Das Internet
im Wahlkampf: Analysen zur Bundestagswahl 2009, eds. E.J. Schweitzer and S. Albrecht. Wiesbaden:
VS Verlag, 136–53.
Marschall, S., and C.K. Schmidt. 2010. “The Impact of Voting Indicators: The Case of the German
Wahl-O-Mat.” In Voting Advice Applications in Europe: The State of the Art, eds. L. Cedroni and D.
Garzia. Napoli: Scripta Web, 65–204.
Marschall, S., and M. Schultze. 2012a. “Normalisierung oder Mobilisierung? Die Auswirkungen
politischer Online-Kommunikation auf die Wahlbeteiligung am Beispiel einer Internet-
Applikation zur Bundestagswahl 2009.” Politische Vierteljahresschrift 53 (3): 444–66.
Marschall, S., and M. Schultze. 2012b. “The Emergence of the ‘Voter 2.0’? VAA Users in a Changing
Political Communication Sphere.” Paper Presented at XXVI Convegno SISP, September 13–15,
Università Roma Tre.
Marschall, S., and M. Schultze. 2012c. “Voting Advice Applications and Their Effect on Voter Turnout:
The Case of the German Wahl-O-Mat.” International Journal of Electronic Governance 5 (3–4):
349–66.
Martin, A. 2012. Young People and Politics: Political Engagement in Anglo-American Democracies. New
York: Routledge.
Mendez, F. 2012. “Matching Voters With Political Parties and Candidates: An Empirical Test of Four
Algorithms.” International Journal of Electronic Governance 5 (3–4): 264–78. [Special Issue: Voting
Advice Applications and the State of the Art: Theory, Practice and Comparative Insights.]
Muthén, L.K., and B.O. Muthén. 2010. Mplus User’s Guide, 6th ed. Los Angeles: CA Muthén & Muthén.
Mykkänen, J., and T. Moring. 2006. “Dealigned Politics Comes of Age? The Effects of Online Candidate
Selectors on Finnish Voters.” Paper at the Conference of Politics on the Internet: New Forms of
Media for Political Action, November 25, Tampere.
Neidhardt, F. 1994. “Öffentlichkeit, öffentliche Meinung, soziale Bewegungen.” In Öffentlichkeit,
öffentliche Meinung, soziale Bewegungen, ed. F. Neidhardt. Opladen: Westdeutscher Verlag, 7–41.
Norris, P. 2000. A Virtuous Circle? Political Communications in Post-Industrial Democracies. Cambridge:
Cambridge University Press.
Owen, D. 2006. “The Internet and Youth Civic Engagement in the United States.” In The Internet and
Politics: Citizens, Voters and Activists, eds. S. Oates, D. Owen, and R.K. Gibson. New York:
Routledge, 17–33.
Pappi, F.U., and J. Brandenburg. 2012. “Die Politikvorschläge der Bundestagsparteien aus Wählersicht.
Zur Konstruierbarkeit von Politikräumen für das deutsche Fünfparteiensystem.” In Wählen in
Deutschland, ed. R. Schmitt-Beck. Baden-Baden: Nomos, 276–301.
Pianzola, J. 2011. “The Interplay Between Political Knowledge and VAA Use and Its Effect on Voting
Behavior.” Paper presented at the Conference “Etmaal 2011,” June 9–10, Amsterdam.
Pianzola, J., A.H. Trechsel, G. Schwerdt, K. Vassil, and M.R. Alvarez. 2012. “The Effect of Voting
Advice Applications (VAAs) on Political Preferences. Evidence from a Randomized Field
Experiment.” Paper presented at 2012 Annual Meeting of the American Political Science
Association, August 30–September 2, New Orleans.
Roßteutscher, S. 2012. “Die konfessionell-religiöse Konfliktlinie zwischen Säkularisierung und
Mobilisierung.” In Wählen in Deutschland, ed. R. Schmitt-Beck. Baden-Baden: Nomos, 111–33.
Schultze: Effects of VAAs on Political Knowledge About Party Positions 67
Appendix: Result Screen of the Wahl-O-Mat for the 2009 German Federal
Election