Framing effects in consumer expectations surveys

In a randomized experiment embedded in a survey, I test the effects of variations in question wording and format on consumer response behavior and the corresponding inflation expectations. To this end, survey participants from a representative sample of German consumers are broken down into four treatment groups and presented with different versions of a question asking for their subjective distribution for inflation over the next 12 months. As part of the experiment, two competing wordings, previously known from leading consumer surveys, are considered: (i) the change in prices in general or (ii) the inflation rate. In addition, I compare the responses to a question asking for consumers' probabilistic beliefs about future inflation, to those from a simpler one asking for the expected minimum, maximum, and most likely inflation rate over the short term. I find that response behavior varies strongly with framing. Simpler wording such as 'prices in general' and a less restrictive format lead to higher mean expected inflation, on average. While simpler wording increases individual uncertainty derived from the subjective histograms, asking for minimum, maximum and mode leads to lower uncertainty about expected inflation.