UC Berkeley

In one large-scale experiment using U.S. respondents on Mechanical Turk (N = 2,899), we studied how subtle differences in framing and context impacted estimates of the Black–White wealth gap. Across our 10 different experimental manipulations of framing and context, respondents consistently overestimated Black family wealth relative to White wealth. There was also substantial variation in the magnitude of these wealth estimates, which ranged from a low of 35 to a high of over 60 percentage points across the conditions. Overestimates were largest when respondents were asked about the Black–White wealth gap at both past and present time points and closest to accuracy when respondents used images as pictorial comparisons for White and Black wealth. Overall, while framing and context certainly affect the magnitude of this misperception, the tendency to overestimate racial wealth equality is extremely robust.