Incumbency Dominance in Letters to the Editor: Field Experimental Evidence

Abstract

This article reports the results of a randomized field experiment conducted three weeks before the 2017 federal election in Germany. Four different versions of a letter to the editor were sent to all the German daily newspapers that handle letters to the editor independently. The versions differed in the subject matter of the letter, the chancellor Angela Merkel versus the main challenger Martin Schulz, and in the evaluation of this subject, positive versus negative. The experiment was designed to test for three different types of media bias: political bias, negativity bias, and incumbency dominance. We find no political bias in the decisions to print letters, and no statistically significant negativity bias. We do observe incumbency dominance: letters about the chancellor were 40% more likely to be printed.