You can't lump "newspapers" together any more than you can lump "reporters" or "people" together.

Different papers have different editorial policies.

Or perhaps the paper just leaves the reporters' words, in which case you'd have to compare different articles by the same reporter. Racism would show up as inconsistent pattern in how the policy's applied or enforced--preferably a fairly strong pattern. If AfAm males were "teens" 40% of the time and whites the same age were "men" 55% of the time, that would strike me as probably random unless we're talking hundreds of instances.

Relying on memory is also a problem. We tend to remember things that stick out, that have our attention drawn to them. Claims of "racism" make things stick out. Often such comparisons are based on comparison with a single story that does the opposite but we remember it as being a strong pattern. I've also seen the condemnation go both ways--AfAm males aged 18 called "teens", it was claimed, denied them adulthood and was tantamount to calling a grown man "boy"; calling an 18-year-old AfAm "man" was to deny him the immaturity that young men have. It really depended on whether one wanted to emphasize or minimize dignity and responsibility, so it's not like there's not some perceiver-side manipulation.