If you don't understand how statistics work, it is easy to be manipulated by intentionally misleading data. Since the races are not evenly represented in our population (64% White, 16% Hispanic, 12% Black, 4% Asian, 2% two or more races), a direct comparison between races in America on any phenomenon will be racially biased and lead to a false conclusion.
Hypothetically, suppose .1% of all Americans have been falsely arrested. Half of them were white and half were black. Saying that 50% of Americans who were falsely arrested were white and 50% were black is true, but it is a racially biased statistic because it doesn’t account for the fact that 64% of Americans are white and only 12% are black.
To factor out this bias, comparisons must be made per 100 black people and per 100 white people (or per 10,000, or per 100,000, or per 1,000,000). By calculating per 100 black Americans, these same data would show that 64% of the people who were falsely arrested were black. And, per 100 white people, only 12% of people who were falsely arrested were white. Now the data lead to an informed conclusion rather than a biased one.
So, the next time you see statistics on race that fail to show the data per 100,000 (or some other round number), you will know that someone is trying to sucker you!