["Oxford Bulletin of Economics and Statistics, EarlyView. ", "\nABSTRACT\nWe study how both who we talk about when we talk about inequality and how we talk about it shape redistributive preferences. In a nationally representative survey experiment in Germany with a 1‐year follow‐up, we exogenously vary the focus of attention (top 10% vs. bottom 10%) and the mode of delivery (subtle priming vs. factual information about the share of total income earned by that group). We examine effects on perceived inequality, general support for redistribution and concrete tax‐based policy preferences. We find three main results. First, simply directing attention to different ends of the distribution shifts support for redistribution. Second, corrective information about the richest 10% generates larger belief updates and stronger short‐run changes in attitudes than information about the poorest 10%. Third, these effects fade within a year, underscoring the limited durability of one‐off interventions. These findings help reconcile mixed evidence from prior inequality‐information experiments by highlighting the importance of both the focus and the mode of communication in shaping effect size and persistence. We discuss implications for the design of redistribution policy communication and for measuring belief dynamics.\n"]