How Political Systems Manage Their Policy Controversies

1. I did not conduct interviews with blame game actors because I am primarily interested in what blame game participants say and do during the actual blame game. To knowwhat they think would certainly be interesting, but it is not a necessary requirement for the analysis of blame game interactions and their consequences. Moreover, I suspect that interview data would be useless for the most part as interviewees are driven by a very strong urge for social desirability when it comes to explaining their roles and actions during a blame game. Another reason against noncontemporaneous interviews is that memories on tactical issues are likely to fade quickly and that these memories are more likely to be influenced by outcomes than by what calculations interviewees were making at the time (Berry, 2002). 2. This is why I opted against counting the number of newspaper articles during a blame game. The second reason for a qualitative assessment is its flexibility, meaning that salience and proximity can be determined and compared across very different policy controversies. 3. UK newspapers: The Daily Telegraph, The Guardian, The Sun; German newspapers: Die Welt, Süddeutsche Zeitung, Bild; Swiss newspapers: Neue Zürcher Zeitung, Tages-Anzeiger, Blick; US newspapers: The Wall Street Journal, The New York Times, USA Today. I retrieved relevant articles from these newspapers mainly via the Factiva database. I used several controversy-specific keywords in each case to identify all relevant articles in the newspapers selected (see Table A2 in the Appendix). The time span during which I searched for articles starts significantly before the start of the blame game and extends beyond the end of blame game interactions until articles related to the blame game could no longer be found (the concrete time span varies considerably from case to case). Moreover, I occasionally consulted newspaper articles from other outlets in case they could provide further insights into a case. 4. Considering these different newspapers also controls for political parallelism, that is, for a situation in which only one political camp shows strong interest in a controversy while other camps largely ignore