Distortions in the elicitation of economic variables arise frequently. A common problem in household surveys is that reported values exhibit a significant degree of rounding. We interpret rounding as a filter that allows limited information about the relationship of interest to pass. We argue...