Voter flow analysis

The electoral flow analysis is a methodical procedure of electoral research for the calculation of voter migration between political parties and non-voters . You always compare two elections with each other, which means that the voter flows of an election always understand each other in relation to a comparative election.

Streams of voters

The movements of voters between parties are referred to as voter flows . Form a voter stream z. B. all those voters who voted for party A in the 1999 European elections but voted B in the 2004 European elections.

In the voter flow analysis, the group of non-voters is listed on both sides, since it is just as connected to voter flows as between the parties.

method

Aggregate data analyzes or individual data analyzes are used for voter flow analyzes or voter migration analyzes. Voter flow analyzes with the help of aggregate data use official results from electoral districts. On the basis of this data, models are calculated that allow conclusions to be drawn about those migration flows that lie behind the changes between two election results.

The procedure for calculating a voter flow analysis is called multiple regression . This means that on the one hand the current election results of the parties are regressed on those of the previous election. On the other hand, the election result of a party in the current election is related to all party results of the old election. For a voter flow analysis, municipalities are divided into homogeneous groups and regression equations are calculated for each group, from which the entire voter flow analysis is then determined.

The idea behind this method is: If one party is strong in an election (e.g. in 2004) exactly in those parishes where another party was strong in the previous election (e.g. in 1999), it will interprets this as voter migration, d. H. many voters switched between these parties.

Voter flow or voter migration analyzes on the basis of individual data primarily use survey data. The voter migration analyzes published by commercial polling institutes on election evenings are largely based on information from the so-called election day surveys , in which employees of the institute in randomly selected polling stations randomly selected voters after voting according to this, the voting behavior in the last election of the same type (so-called recall question) as well as some other features. In order to record the previous voting behavior of non-voters, surveys are carried out in the week before the election in question, in which - unlike in elections - non-voters also take part. Data from official statistics (age structure) and representative election statistics (voting behavior in age groups) are combined to determine the previous voting behavior of those who have since died, who have moved away and who have left the electorate in another way. From this information and the number of votes, non-voter numbers and so on achieved in the current election, absolute frequencies can be determined for all combinations of previous and current voting behavior. Adding these up for the voters of a particular party in the previous election does not usually result in the number of votes that this party actually achieved. In order to correct this error, iterative adaptation methods are then used, which optimally distribute the error in the data matrix, but do not eliminate it.

advantages

Analyzes of voter flows enable statements to be made about the strength of parties' mobilization in selected elections. The “retention rate” shows how many voters from a party in the previous election have re-elected that party in the current election. Voter analysis shows how the electorate of a new party is made up. Further advantages result from the data source: Voter flow analyzes examine actual behavior instead of collected statements (such as surveys ). There are no problems with sampling errors or incorrect information. Official election results are available for almost every election , so long time series can be calculated using voter flow analyzes.

disadvantage

The results for smaller parties are relatively uncertain. The statistical process is extremely complex. The higher the aggregation of the data (e.g. constituencies compared to electoral districts), the more uncertain the conclusions become, since the problem of the ecological fallacy is more serious.

Voter migration analyzes based on individual data are not confronted with the problem of inferring individual behavior from aggregate data. However, it is uncertain whether the interview information reflects actual voting behavior. This applies in particular to voting behavior in the previous election. In response to the recall question, around 70 percent of respondents in Germany stated the behavior that they stated in interviews at the time of the previous election. Reasons for this error could include: a. Memory problems and the pursuit of cognitive consonance. As a result, with this method, the proportion of swing voters is systematically underestimated.

literature

Uwe Gehring: Voter migration balance sheets of the federal election 1990: A review of the infas concept with data from the research group Elections. In: Hans Rattinger , Oscar W. Gabriel, Wolfgang Jagodzinski (ed.): Elections and political attitudes in a united Germany. Peter Lang, Frankfurt / Main 1994, pp. 93-112.
Christoph Hofinger , Günther Ogris : Oracles of the Modern Era: What do election exchanges, voter flow analyzes and election day projections achieve? In: Austrian Journal for Political Science 31, 2002/2, ISSN 1615-5548 , pp. 143–158.
Richard Hilmer, Michael Kunert: Voter migration: The model of Infratest dimap . In: Jürgen W. Falter , Oscar W. Gabriel , Bernhard Weßels (ed.): Elections and voters. Analyzes on the occasion of the Federal Parliament election 2002 . VS Verlag für Sozialwissenschaften, Wiesbaden 2005, ISBN 3-531-14137-6 , pp. 134–156.
Manfred Küchler: Estimating Voter Migration: New Attempts at Solutions. In: Max Kaase, Hans-Dieter Klingemann (Hrsg.): Elections and the political system. Analyzes on the occasion of the federal election in 1980. Westdeutscher Verlag, Opladen 1983, pp. 632–651.
Claus Laemmerhold: Bend and break: The non-voters in the Procrustal bed of migration balances. In: Max Kaase, Hans-Dieter Klingemann (Hrsg.): Elections and the political system. Analyzes on the occasion of the federal election in 1980 . Westdeutscher Verlag, Opladen 1983, pp. 624-631.
Harald Schoen: Voter change and change of election. A comparative study. VS Verlag für Sozialwissenschaften, Wiesbaden 2003.
Harald Schoen: Change of choice. In: Jürgen W. Falter , Harald Schoen (Hrsg.): Handbuch Wahlforschung . VS Verlag für Sozialwissenschaften, Wiesbaden 2005, pp. 367–387.
Harald Schoen: Does ticket-splitting decrease the accuracy of recalled previous voting? Evidence from three German panel surveys. In: Electoral Studies 30, 2011, pp. 358–365.

Web links

Equation of a voter flow analysis

Individual evidence

↑ Uwe Gehring: Voter migration balance sheets of the federal election 1990: A review of the infas concept with data from the research group Elections. In: Hans Rattinger, Oscar W. Gabriel, Wolfgang Jagodzinski (ed.): Elections and political attitudes in a united Germany. Peter Lang, Frankfurt / Main 1994, pp. 93-112; Richard Hilmer, Michael Kunert: Voter migration: The model of Infratest dimap . In: Jürgen W. Falter , Oscar W. Gabriel , Bernhard Weßels (ed.): Elections and voters. Analyzes on the occasion of the Federal Parliament election 2002 . VS Verlag für Sozialwissenschaften, Wiesbaden 2005, pp. 134–156; Manfred Küchler: Estimating Voter Migration: New Attempts at Solutions. In: Max Kaase, Hans-Dieter Klingemann (Hrsg.): Elections and the political system. Analyzes on the occasion of the federal election in 1980. Westdeutscher Verlag, Opladen 1983, pp. 632–651, Claus Laemmerhold: Auf Biegen und Brechen: The non-voters in the Procrustal bed of migration balances. In: Max Kaase, Hans-Dieter Klingemann (Hrsg.): Elections and the political system. Analyzes on the occasion of the federal election in 1980 . Westdeutscher Verlag, Opladen 1983, pp. 624-631.
↑ See Gary King: A Solution to the Ecological Inference Problem. Princeton University Press, Princeton 1997.
↑ Harald Schoen: Does ticket-splitting decrease the accuracy of recalled previous voting? Evidence from three German panel surveys. In: Electoral Studies 30, 2011, pp. 358–365.
↑ Where the voters go, nobody knows (not even the ARD) ›Krautreporter. Retrieved October 30, 2019 .

[1] Uwe Gehring: Voter migration balance sheets of the federal election 1990: A review of the infas concept with data from the research group Elections. In: Hans Rattinger, Oscar W. Gabriel, Wolfgang Jagodzinski (ed.): Elections and political attitudes in a united Germany. Peter Lang, Frankfurt / Main 1994, pp. 93-112; Richard Hilmer, Michael Kunert: Voter migration: The model of Infratest dimap . In: Jürgen W. Falter , Oscar W. Gabriel , Bernhard Weßels (ed.): Elections and voters. Analyzes on the occasion of the Federal Parliament election 2002 . VS Verlag für Sozialwissenschaften, Wiesbaden 2005, pp. 134–156; Manfred Küchler: Estimating Voter Migration: New Attempts at Solutions. In: Max Kaase, Hans-Dieter Klingemann (Hrsg.): Elections and the political system. Analyzes on the occasion of the federal election in 1980. Westdeutscher Verlag, Opladen 1983, pp. 632–651, Claus Laemmerhold: Auf Biegen und Brechen: The non-voters in the Procrustal bed of migration balances. In: Max Kaase, Hans-Dieter Klingemann (Hrsg.): Elections and the political system. Analyzes on the occasion of the federal election in 1980 . Westdeutscher Verlag, Opladen 1983, pp. 624-631.

[2] See Gary King: A Solution to the Ecological Inference Problem. Princeton University Press, Princeton 1997.

[3] Harald Schoen: Does ticket-splitting decrease the accuracy of recalled previous voting? Evidence from three German panel surveys. In: Electoral Studies 30, 2011, pp. 358–365.

[4] Where the voters go, nobody knows (not even the ARD) ›Krautreporter. Retrieved October 30, 2019 .