Publications

Recent polling failures highlight that election polls are prone to biases that the margin of error customarily reported with polls does not capture. However, such systematic errors are difficult to assess against the background noise of sampling variance. Shirani-Mehr et al. (2018) developed a hierarchical Bayesian model to disentangle random and systematic errors in poll estimates of two-party vote shares at the election level. The method can inform realistic assessments of poll accuracy. We adapt the model to multiparty elections and improve its temporal flexibility. We then estimate bias and variance in 5,240 German national election polls, 1994–2021. Our analysis suggests that the average absolute election-day bias per party was about 1.5 percentage points, ranging from 0.9 for the Greens to 3.2 for the Christian Democrats. The estimated variance is, on average, about twice as large as that implied by usual margins of error. We find little evidence of house or mode effects. Common biases indicate industry effects due to similar methodological problems. The Supplementary Material provides additional results for 1,751 regional election polls.

Directed acyclic graphs (DAGs) are now a popular tool to inform causal inferences. We discuss how DAGs can also be used to encode theoretical assumptions about nonprobability samples and survey nonresponse and to determine whether population quantities including conditional distributions and regressions can be identified. We describe sources of bias and assumptions for eliminating it in various selection scenarios. We then introduce and analyze graphical representations of multiple selection stages in the data collection process, and highlight the strong assumptions implicit in using only design weights. Furthermore, we show that the common practice of selecting adjustment variables based on correlations with sample selection and outcome variables of interest is ill-justified and that nonresponse weighting when the interest is in causal inference may come at severe costs. Finally, we identify further areas for survey methodology research that can benefit from advances in causal graph theory.

Digital contact tracing apps have been introduced globally as an instrument to contain the COVID-19 pandemic. Yet, privacy by design impedes both the evaluation of these tools and the deployment of evidence-based interventions to stimulate uptake. We combine an online panel survey with mobile tracking data to measure the actual usage of Germany's official contact tracing app and reveal higher uptake rates among respondents with an increased risk of severe illness, but lower rates among those with a heightened risk of exposure to COVID-19. Using a randomized intervention, we show that informative and motivational video messages have very limited effect on uptake. However, findings from a second intervention suggest that even small monetary incentives can strongly increase uptake and help make digital contact tracing a more effective tool.

Self-reports of adherence to non-pharmaceutical interventions in surveys may be subject to social desirability bias. Existing questioning techniques to reduce bias are rarely used to monitor adherence. We conducted a list experiment to elicit truthful answers to the question whether respondents met friends or acquaintances and thus disregarded the social distancing norm. Our empirical findings are mixed. Using the list experiment, we estimate the prevalence of non-compliant behavior at 28%, whereas the estimate from a direct question is 22%. However, a more permissively phrased direct question included later in the survey yields an estimate of 47%. All three estimates vary consistently across social groups. Interestingly, only the list experiment reveals somewhat higher non-compliance rates among the highly educated compared to those with lower education, yet the variance of the list estimates is considerably higher. We conclude that the list experiment compared unfavorably to simpler direct measurements in our case.

We present a polling strategy to predict and analyze runoff elections using the 2017 French presidential race as an empirical case. This strategy employs rejective probability sampling to identify a small sample of polling stations that is balanced with respect to past election results. We then survey the voters’ candidate evaluations in first-round exit polls. We poststratify the voter sample to first-round election returns to account for nonresponse and coverage issues, and impute missing candidate evaluations to emulate campaign learning. Next, the votes for eliminated competitors are redistributed according to their supporters’ lower-order preferences. Finally, the predictions are validated against official results and other polls. We end with a discussion of the advantages and limitations of this approach.

dc.title:

Research on parliamentary careers has paid little attention to variations in pre-parliamentary career patterns and their value in explaining legislators’ parliamentary success. Using sequence and cluster analysis, this article identifies typical career tracks taken by Party Animals, Local Heroes, Late Bloomers, Land Legislators, High-Flyers and Career Changers based on a comprehensive dataset of German parliamentarians’ biographies (1998–2014). The analysis confirms the role of the party as the primary career facilitator before and within parliament. Nonetheless both Career Changers and High-Flyers climb the greasy pole all the way to the national parliament without much service to the party. The former type, however, suffers from a lack of networks and experience, which is reflected in the limited career success within parliament. This article demonstrates that the use of sequence analysis on career paths offers a promising approach in distinguishing and explaining the opportunities, choices and obstacles MPs face in parliament.

Does the opportunity to deliver visual instead of verbal stimuli of political knowledge to respondents in web-based surveys make a difference in terms of data quality? For instance, does the presentation of visual knowledge items reduce cheating, that is, looking up the answer via the Web? And do visual and verbal stimuli capture the same underlying construct? To test whether the use of visuals to measure political knowledge effectively makes a difference, we administer a question form experiment in an online survey of the German Longitudinal Election Study. Respondents are randomly assigned to one of two question formats—visual or verbal—and are asked to solve a set of eight questions on political leaders and their offices. The instruments are validated based on nonparametric item response theory and analyses of response latency. While there is no clear evidence for cheating behavior under either of the conditions, both instruments form strong knowledge scales. Results from a regression analysis indicate that the scales provide measures of closely related but not identical concepts.

This article reports on an attempt to forecast the outcome of the 2013 election to the German Bundestag. In contrast to the predominant academic approach to forecast incumbent vote shares from measures of government popularity, economic conditions and other fundamental variables, we entirely relied on data from published trial heat polls. Opposite to common practice in the news media, we did not take isolated polls as election forecasts in their own right. Instead, we used historical data to assess empirically the relationship between polls and election outcomes, and combined extrapolations from current polls in a Bayesian manner. The forecast was published one month ahead of the election. The retrospective evaluation of our method was added after the election. While our method is parsimonious and provides a large lead time, the performance at the 2013 election was underwhelming. We offer additional suggestions how the approach can be improved in future scenarios.

Speculations about whether strategic voting made a difference to the outcome of an election regularly whip up the passions of pundits, party strategists, electoral reformers and scholars alike. Yet, research on strategic voting's political effect has been hampered by the scarcity of data on district level party preferences. We propose the use of Bayesian small area estimation to predict district level preferences from just a handful of survey responses per district and comparing these predictions against election results to estimate how many voters switched sides in each district. We apply the approach to estimate how many seats changed hands as a result of strategic voting at the 1997 and 2001 UK general elections. Despite similar rates of strategic voting in both elections, the number of seats that were affected was markedly greater in 1997. Interestingly, the Liberal Democrats turn out to win the most seats because of strategic voting. We also estimate how many votes went in the ‘wrong’ direction—away from otherwise viable candidates. We validate our results by using journalistic sources and compare them with previous published estimates.

In their seminal paper, Carey and Shugart (1995) argue that electoral rules predetermine whether election campaigns are predominantly characterized by competition between parties, or among candidates within parties. In particular, electoral systems that provide weak party control over ballot access, that allow voters to express a preference vote for individual candidates, and that select candidates on the basis of the votes they earn individually are supposed to offer strong incentives for candidates to cultivate a personal vote. In this paper, we investigate Carey and Shugart's original claims about the instrumental motivations of candidates in open-ballot proportional representation (PR) elections using survey data and election statistics from the 2007 Swiss National Council elections. As opposed to previous work, we develop and employ more direct and valid measures of intra- and interparty incentives to wage personal campaigns. Our empirical results suggest that the candidates' campaign focus and expenditures are closely linked to intraparty competition, but not interparty competition. Previously used proxies of intraparty competition, such as district magnitude or the number of candidates on a party list (or ratios thereof), turn out to be but remotely related to the candidates' campaign behavior.

We illustrate the power of “logical models” (Taagepera, 2007) by offering a three-parameter model of the relationship between the effective number of parties and electoral turnout that makes use of the constraints on what parameter values are internally coherent given boundary conditions to specify functional form, and seeks not optimal curve fitting but rather a direct model testing. In our model, one parameter reflects an effect that generally acts to increase turnout as the effective number of parties increases, another an effect that generally acts to decrease turnout as the effective number of parties increases, while a third parameter allows for baseline variation in turnout across countries (or within countries across elections). We fit this model to district-level data from 237 elections held in 17 countries, representing a wide range of electoral system types generating multi-party contests, with over 20,000 district-election observations. The basic intuition, that turnout rises to a peak as the effective number of parties increases and then falls slowly, fits our data pretty well.

Figures from postelection surveys often grossly overestimate election turnout. Two distinct phenomena are responsible for this gap: overrepresentation of actual voters and vote misreporting by actual nonvoters among survey respondents. Previous accounts of turnout bias are inconclusive in that they either focus on a single component, or fail to separate between the two. In this paper, we formally decompose turnout bias in election surveys into its constituent parts, assess their empirical prevalence and heterogeneity using an extensive collection of 49 vote validation studies from six countries, and employ Bayesian meta regression techniques to account for cross-study differences. Our results indicate that both election and survey characteristics such as actual voter turnout and survey response rates differentially affect the components of turnout bias. We conclude with a discussion of the threats and potentials of our findings for survey-based comparative electoral research.

We draw attention to a simple yet underappreciated way of forecasting the outcomes of elections involving two rounds of voting: surveying the voters’ candidate evaluations in first round exit polls, poststratifying the sample proportions of reported votes to official first round election returns, and redistributing the votes for eliminated competitors according to their supporters’ lower-order preferences among the viable alternatives in round two. We argue that the approach is likely to outperform standard pre-election surveys, due to its better coverage and reduced measurement error, and the possibility of correcting for sample selection. We set out the practical details of the method and demonstrate its usefulness by employing a recent German mayoral election as an empirical case. Thirteen candidates were competing in the first round, while there were six candidates in the decisive second round. The runoff result was forecast two weeks in advance with an average absolute error of less than one percentage point.

How quickly, to what extent and under what conditions do voters and elites adapt to new electoral institutions in order to not waste their votes and effort on hopeless competitors? A latent-curve model of strategic adaptation is developed and fitted to district-level election data from Spain. The extent of strategic adaptation is generally found to vary with the strength of the electoral system. However, grave ethnic tensions are demonstrated to seriously retard adaptation even under favourable institutional conditions.

Measures of constituency preferences are of vital importance for the study of political representation and other research areas. Yet, such measures are often difficult to obtain. Previous survey-based estimates frequently lack precision and coverage due to small samples, rely on questionable assumptions or require detailed auxiliary information about the constituencies' population characteristics. We propose an alternative Bayesian hierarchical approach that exploits minimal geographic information readily available from digitalized constituency maps. If at hand, social background data are easily integrated. To validate the method, we use national polls and district-level results from the 2009 German Bundestag election, an empirical case for which detailed structural information is missing.

The Swiss party system has become strongly polarized over the last decade, following the rise of the Swiss People's Party and the electoral losses of center parties. This article suggests that these developments are, at least in part, a consequence of strategic behaviour among voters. As the government policy is the result of institutionalized multiparty bargaining, voters have incentives to compensate for this watering-down by supporting parties whose positions are more extreme than their own. This article empirically tests extent and conditions of compensatory voting in the 2007 National Council Elections using Selects survey data. Our results suggest that compensatory voting generally outweighs voting based on ideological proximity and increases with rising district magnitude.

Blais (2006) and Blais and Aarts (2006) in their review essays on voter turnout call attention to a striking puzzle about the link between electoral systems and turnout, namely that, ceteris paribus, proportional representation (PR) systems with many parties appear to have higher national-level turnout than single-member district (SMD) plurality systems with few parties, yet turnout does not increase with the (effective) number of parties (ENP) at the national level. To address this puzzle we turn to district-specific within-nation panel data from Switzerland and Spain. Our country-specific findings allow us to explain the national-level puzzle as essentially an ecological artefact, in that the multi-member districts found in proportional systems, on average, do exhibit higher turnout than SMDs, but turnout does not rise with district magnitude, m, once we move beyond the contrast between m = 1 and m > 1. Using a more sophisticated approach to measuring political competition that does not treat all PR systems as generating identical turnout incentives (Grofman and Selb, 2009), we seek to explain this puzzle by showing both empirically and theoretically that (1) proportionality does not necessarily increase with district magnitude, and (2) competition does not necessarily increase with district magnitude.

Ever since Lazarsfeld and his colleagues (1944) seminal study, it has become common wisdom that election campaigns, if anything, serve the activation of voters fundamental predispositions. However, disagreement emerges on the role of partisan orientations. Although some authors consider them as fundamental predispositions, which are activated during the campaign and subsequently act as filters for incoming information, others argue that party attachments are simple running tallies of political assessments, which are constantly updated in response to campaign events, or decision shortcuts for voters innocent of substantial information. In this study, we scrutinize the role of partisan orientations in a direct-democratic campaign using data from a panel survey fielded during the run-up to the 2006 Swiss asylum law referendum. We find that, as voters accumulate knowledge in the course of the campaign, vote intentions dramatically converge on pre-campaign partisan orientations. Moreover, voters, whose earlier issuespecific and partisan orientations collide, tend to resolve their ambivalence in favour of their partisan leanings. These results corroborate the view of partisanship as a fundamental predisposition.

As Selb lA deeper look at the proportionality-turnout nexus. Comparative Political Studies. forthcoming) observes. the standard measure of (posterior) competitiveness in elections used in plurality contests, namely the difference between the winner and the second place finisher. has no "self-evident counterpart in multi-member PR districts". Following lines similar to Taagepera and Grofman [2003. Mapping the indices of seats-votes disproportionality and inter-election volatility. Party Politics 9. 659-677] we identifY six properties that any index of competition should satisfY and then offer a new measure that satisfies all six properties which is applicable to virtually any electoral rule. We then apply our new and more fully general index of political competition to data from PR elections in Switzerland to see the extent to which the index of competition is correlated with levels of voter turnout.

Compulsory voting (CV) undoubtedly raises electoral turnout. Yet does it also affect individual party choices and aggregate election outcomes? Previous studies have focused on partisan or 'directional' effects of CV in favour of, for example, social-democratic or anti-establishment parties. These effects are usually small, however. Using survey data from the Belgian General Elections Study, this article finds that CV primarily affects the consistency, rather than the direction, of party choices. In particular, the analyses suggest that CV compels a substantial share of uninterested and less knowledgeable voters to the polls. These voters, in turn, cast votes that are clearly less consistent with their own political preferences than those of the more informed and motivated voluntary voters. Claims that CV promotes equal representation of political interests are therefore questionable.

Previous research has identified several structural and situational factors that affect party cohesion in parliamentary voting behaviour. The potential role of leadership has been neglected so far. The authors apply a latent variable approach to model leadership effects in roll call votes from the European Parliament (EP), 1979-2001. Other things being equal, their findings suggest that a small but significant 7 per cent share of the total variance in party group cohesion is due to the party group leaders. About 40 per cent of this leader component can be accounted for by their experience inside the European institutions, their career prospects, and their ideological positions.

Evidence that turnout is higher under proportional representation (PR) than in majoritarian elections is overwhelming. Yet previous research has largely failed to explain why. One line of argument maintains that higher turnout under PR is a by-product of larger party systems. However, a larger number of parties has been demonstrated to depress turnout. Alternatively, it is argued that majoritarian electoral systems tend to produce safe seats and that voters have little incentive to turn out there. Thus, uneven turnout over electoral districts due to variable intensities of local competition is made responsible for the lower overall turnout. Empirical evidence on this conjecture is scant. This article scrutinizes the relationship between electoral rules, competition, and turnout with district-level data from 31 national elections. Results from a heteroscedastic model indicate that lower net turnout in majoritarian systems is indeed a consequence of uneven turnout over districts due to variable levels of local competitiveness.

Extensive research efforts notwithstanding, scholars continue to disagree on the nature and meaning of party identification. Traditionalists conceive of partisanship as a largely affective attachment to a political party that emerges in childhood through parental influences and tends to persist throughout life. The revisionist conception of partisanship is that of a running tally of party utilities that is updated based on current party performance. We attempt to reconcile both schools of thought in an individual difference perspective, showing that the party loyalties acquired through parental influences confirm better the traditional view, while the attachments of individuals who did not inherit their parents' party loyalties exhibit features more closely matching the revisionist predictions. The analysis is facilitated by uniquely suited longitudinal household data emanating from the German Socio-Economic Panel Study that allow to study party identifications of young adults and their parents on an annual basis from 1984 to 2007.

Studies of the emergence and electoral success of new parties frequently suffer from conceptual and methodological flaws: effects of electoral institutions that primarily operate at the level of the constituencies are specified at the national level; central explanatory variables such as electoral demands and competitors’ responses are measured with crude proxies; the interplay of formation and success is overlooked; and empirical models neglect the point that the underlying data may exhibit dependencies both in space and time. In this article, we highlight these problems and offer some potential solutions using the Swiss Green Party (GPS) as an empirical case.

Voters in polities that make heavy use of direct democracy are frequently confronted with ballots that contain a multitude of propositions. Claims that direct legislation elections overwhelm voters with choices they are not competent to make should particularly apply to such demanding settings. Yet, evidence on the effects of lengthy ballots on voting behavior is scant. This study reviews theories of decision-making under uncertainty, and tests their predictions in a mixed heteroscedastic model of vote choice that is fitted to a unique collection of survey and contextual data on Swiss referendums. Increasing ballot length is demonstrated to interfere with the voters’ ability to translate their political preferences into consistent policy choices.

Why do election results at national and regional parliamentary elections in Switzerland differ so widely? And why are these differences more pronounced in some constituencies than in others? This study discusses competing theoretical views of the linkage between elections held at multiple federal levels, and empirically tests their predictions using official election statistics and contextual data from Swiss national and cantonal elections between 1999 and 2003. Despite the spatially and temporally limited scope of this analysis, one conclusion suggests itself: current theories of the linkage suffer from their neglect of features of the electoral systems which may vary between different types of elections. Taking these institutional variations into account, we find a strong systematic relationship between election outcomes at different levels. Moreover, the linkage of election outcomes is, to some extent, contingent upon the degree to which regions are integrated into the national political system: while national trends in party support tend to drive election outcomes in nationally well-integrated cantons, election results ostensibly follow regional electoral developments in more peripheral cantons.