The National Crime Victimization Survey, which is administered twice a year to large national random samples of US residents, collects data from representative samples of people of all racial categories about any and all crimes they have been victimized by, regardless of whether they reported it, or if there were any arrests.
They collect as much detail as possible about the nature of the crime and the victim's report of the suspects (number of suspects, race, gender, etc.)
The results show the same racial disparities in offenders as the official arrest data does, and it holds for victims of all races. IOW, it isn't just white victims claiming the offender was black. Overall, close to 60% of crime is within-race, so a victim is more likely to say their offender was the same race as them than any other single race. However, the differences in these ratios shows that blacks are disproportionately more likely to victimize not only their own race but all other races. Black violent crime victims are 6 times more likely to say their offender was Black than White, and 10 time more likely to be Black than Hispanic. But White victims are only 4 times as likely to say their offender was White than Black, and 5 times more likely White than Hispanic. While Hispanic victims are only twice as likely to say their offender was Hispanic than Black and that Blacks and Whites were equally likely to be the offender. This pattern is the exact opposite of what would be predicted by the total proportion of each group in the population, showing that crime victims within each racial group report a % of black offenders that is far higher than their % of the population. Overall, the data show that blacks are 3 times as likely as whites and twice as likely as Hispanics to be identified as the offender in a violent crime by crime victims. And this pattern holds for both male and female victims and victims of all age groups.
And those relative rates of engaging in violent crime map on very closely to the relative likelihood of being shot by the cops, which is also about 3 times as likely for blacks than whites, and twice as likely for blacks than Hispanics.
Also, because about 60% of violent crime is within racial group, greater rates of crimes committed by blacks predicts greater victimization rates among blacks. That is just what the data shows. Relative to their proportion of the total population, blacks are about 50% more likely than whites to be crime victims overall, and 50% more likely to be violently victimized by a member of their own race.
Finally, according to victims of all races, their offenders were more likely to have a weapon in general and a especially a firearm when the offender was black.
These data (with highly similar patterns in every other year) provide strong independent confirmation of the racial disparities shown in police arrest reports, with actual greater propensity to commit violent crimes (and more likely to use a gun when doing so) among blacks in the US, than among either whites or Hispanics.
These facts contradict the unsupported speculation that official crime rates based on arrests are merely due to racial bias by police, and rather provide independent evidence that those rates largely reflect the greater rates of actual crimes committed by blacks. It is important to note that the objective reality of greater actual crime rates does NOT support any notions that the source of these disparities in committing crimes lies in any biological/genetic/innate racial differences. 100% could plausibly be due to differences in SES combined with differences in cultural/political history of these groups.
But none of that changes the facts relevant to who/why the cops shoot, which are that 1) Cops shoot suspects more likely to be engaged in violent crime and who are encountered in high violent crime areas, 2) different regions (e.g., urban, suburban, rural) differ in racial disparities of crime rates and this strongly predicts variations in who is more likely to be shot, and 3) On the national average, blacks are more likely to engage in violent crime and be in high violent crime areas. All of which combined logically imply that blacks in the US are more likely to be shot by the police, even if the cops respond with zero racial bias in how they respond to the objective facts of a situation. Which of course does not mean that no cops ever respond with deadly force due to racial bias, but that such bias is not evidenced by statistical racial disparities in police shootings b/c those disparities are almost entirely accounted for by disparities in violent crime rates, even when crime rates are determined by victim accounts whether reported to the police or not, and regardless of the victim's own race.
