Three questions for… Martin Haselmayer


Martin Haselmayer researches negative communication in politics and the media and its effects on democratic competition. He and a colleague have created a sentiment lexicon to analyse "negative campaigning", for example in Austrian National Council elections. AUSSDA asked him about the results of his research.

What is the study about?

The data serve as a basis for the analysis of negative political and media communication and political mediation. It is about empirically analysing "how negatively" parties communicate, criticise each other, or media report on politics. While previous research has consistently relied on a simple, dichotomous distinction between "positive" and "negative" communication, this approach argues that the intensity or tonality of negative statements has consequences for the reception and effects of negativity. In doing so, we draw on experimental research that shows that the strength of negative statements influences the perception of recipients. On the one hand, criticism is an essential component of political competition and fulfils indispensable functions in democratic systems. It serves, for example, to delineate positions and contents or to point out misconduct and weaknesses and can thus contribute to ensuring political accountability. On the other hand, studies show that violent attacks or particularly offensive language can demobilise voters or trigger cynicism towards politics and democracy or contribute to the polarisation of society. It is therefore important to work out nuances in political communication in order to distinguish different consequences from each other.

To this end, my colleague, Marcelo Jenny, and I first developed tools for (semi-)automated sentiment analysis of German-language texts, which enable an analysis of the intensity of negative communication. For this purpose, we had the negativity of sentences from press releases and parliamentary debates rated several times on a five-part scale by means of online crowdcoding. These data are available as training data for machine learning approaches and form the basis for a German-language sentiment dictionary.

In our research, we use the crowd-coded data directly to analyse "negative campaigning" in Austrian national elections. In addition to analysing party communication, we also used the sentiment lexicon to investigate the media communication of politics in election campaigns.

Furthermore, we used the training data for machine learning-based analyses of the negativity of parliamentary debates in the National Council in an interdisciplinary project.

What is the most exciting aspect of the study for you? Were there any surprising results?

The starting point of our research was the conceptual mismatch between the measurement of "negative campaigning" in existing research and the public and media perception of this campaign strategy. Previous research based on our data confirms that parties show different patterns in terms of the volume of negative communication and the intensity or tonality of their attacks. For example, although coalition partners often criticise each other in election campaigns, this criticism is usually less intense than negative campaigning between government and opposition parties. This may indicate that while these parties are fighting out substantive differences in the election campaign, they are at the same time careful not to damage future coalition options by smashing china verbally. A very recent result of our research shows that the speeches of female MPs in the National Council are less negative than those of their male colleagues. Women MPs can also have a positive influence on their social environment: Negativity in debates and in parliamentary groups decreases when the proportion of female MPs is higher.

Why did you decide to make the data freely available?

This decision seems to me to have no alternative: Scientific research thrives on the free availability of data and analytical tools. An essential aspect is to subject existing results and approaches to constant critical consideration and review. We also hope that other researchers will use this data to answer exciting research questions, shedding light on the understanding of negative political communication, media mediation of politics and its consequences on democratic competition and voters from different perspectives.

  • Martin Haselmayer is a postdoctoral researcher at the Varieties of Egalitarianism (VoE) project at the Department of Politics and Public Administration (University of Konstanz) and a senior research fellow at the Department of Government (University of Vienna). Previously, he was a pre- and postdoctoral researcher in the Austrian National Election Survey (AUTNES). He studied political science and Romance studies in Vienna, Aix-en-Provence and Lyon. His research interests are political communication, party competition and text analysis.
Auf dem Foto ist Dr. Martin Haselmayer zu sehen.

Martin Haselmayer (Foto: privat)