Group Appeals of Parties in Times of Economic and Identity Conflicts and Realignment

Party–group relations are today characterized by various forms of alignments. These include the persistence of traditional class alignments, the realignment of economic groups due to identity politics and alignments of groups at the centre of identity politics. This study analyses the group-based messaging of parties in relation to these three groups. We argue that, contrary to the catch-all party thesis, positive and negative group appeals of parties are (still) structured by parties’ support among social groups and by groups either liked or disliked by their voters. Our argument is tested through a content analysis of election materials in Germany, Switzerland and the Netherlands, combined with survey data. The results indicate that group appeals are indeed shaped by electoral support among social groups and attitudes towards them. Results also suggest that the former effect is present for traditionally aligned economic and identity politics groups, but not for realigned economic groups.

For the Netherlands and Germany, we analysed the TV ads which parties ran in the stateallocated free time slots on the public TV channels.Coding involved transcribing the audio of the ad and the visual content was left out of the analysis.Each ad was coded only once, irrespective of the number of times it was aired, given that this information was not available.
For Switzerland, newspaper ads were collected from three types of newspapers in the Germanand French-speaking parts of the country.The selection includes: two quality newspapers (NZZ, Le Temps), two regional newspapers (Tagesanzeiger, Tribune de Genève) and two tabloids (Blick, Le Matin).For newspapers it was possible to code the relative importance of particular ads by coding each ad, even if it already appeared in another newspaper or on another day.This allows for a more accurate description of the emphasis political parties put on different messages for the Swiss case.

C) Additional information on the coding of group appeals
The original coding of group appeals was based on a variable with 55 groups, which were each listed twice to distinguish between positive and negative references (see Table A10).

Table A3
Regression Table: Traditionally aligned economic group categories

Table A4 Regression
Table: Realigned economic group categories Note: The dependent variable is the number of group appeals to one of the traditionally aligned group categories (large employers, small employers, farmers, employees) or to one of them with attitudinal data (large employers, employees).Cell entries are exponentiated regression coefficients (Incident Rate Ratios) from negative binomial regressions with robust standard errors (clustered by party) in parentheses.The number of core sentences is included as exposure variable.In models that include voter's attitudes, the number of observations is reduced because those data are not available for the German Radical Right.*** p<0.001, ** p<0.01, * p<0.05 Note: The dependent variable is the number of group appeals to one of the realigned group categories (workers, poor people, socio-cultural prof., students) or to one of them with attitudinal data (workers, poor people).Cell entries are exponentiated regression coefficients (Incident Rate Ratios) from negative binomial regressions with robust standard errors (clustered by party) in parentheses.The number of core sentences is included as exposure variable.In models that include voter's attitudes, the number of observations is reduced because those data are not available for the German Radical Right.*** p<0.001, ** p<0.01, * p<0.05

Table A5
Regression Table: IP group categories Note: The dependent variable is the number of group appeals to one of the IP group categories (immigrants, LGBT people, women).Cell entries are exponentiated regression coefficients (Incident Rate Ratios) from negative binomial regressions with robust standard errors (clustered by party) in parentheses.The number of core sentences is included as exposure variable.In models that include voter's attitudes, the number of observations is reduced because those data are not available for the German Radical Right.*** p<0.001, ** p<0.01, * p<0.05

Table A6
Regression Tables: Reference logic Note: The dependent variable is the number of negative group appeals to one of the following group categories: large employers, rich people, immigrants.Cell entries are exponentiated regression coefficients (Incident Rate Ratios) from negative binomial regressions with robust standard errors (clustered by party) in parentheses.The number of core sentences is included as exposure variable.In models that include voter's attitudes, the number of observations is reduced because those data are not available for the German Radical Right.*** p<0.001, ** p<0.01, * p<0.05 Note: The dependent variable is the number of positive group appeals to one of the following group categories: employees, large employers, immigrants, poor people, LGBT people, women, workers.Cell entries are exponentiated regression coefficients (Incident Rate Ratios) from negative binomial regressions with robust standard errors (clustered by party) in parentheses.The number of core sentences is included as exposure variable.In models that include voter's attitudes, the number of observations is reduced because those data are not available for the German Radical Right.*

Table A9
Number of press releasesPress releases were collected for a period of two months prior to the election date.Only press releases with substantial political content are included in the corpus, excluding announcements of upcoming events, for example.

Table A10
List of original group categories