Yeah, “low-income” has a few different definitions across different agencies and years for different purposes. Sometimes it’s a multiple of a “poverty threshold” (e.g 200% of the poverty threshold) and sometimes it’s relative to median income in the area (e.g. 80% of the median family income for the area).

This survey seems to be inline with the latter. When averaging across occupational averages, median and mean are not as egregiously different as median vs mean household income as outliers are sort of accounted for across groups (kinda). But I agree with your sentiment.