Linking to actual sources would reveal that the keywords the IRS was looking for were politically biased, yes, but across the spectrum. The keywords included "Tea Party", "Patriot", "Progressive", and "Occupy." https://www.npr.org/2017/10/05/555975207/as-irs-targeted-tea...

"Biased but across the spectrum" is nonsense.

Purely semantic arguments aren't helpful to anyone.

The word "bias" clearly has two senses in this context. The original term from signal processing indicates a persistent offset, which got appropriated in politics to reflect the idea of a "lean" in coverage. So now "Bias" means "politically charged in some direction or another".

So you can have a "biased" term ("occupy") next to another biased term ("tea party") in a search. And it's reasonable to call the whole thing a collection of biased terms even though by the original definition I guess you'd say they cancel out and are "unbiased".

Language is language. It may not be rational but it's by definition never "nonsense". Don't argue with it except to clarify.

> Your comment is longer nonsense.

Sigh, here we go.

> Individual data points in a population cannot be biased.

Indeed they[1] cannot! By the first definition I listed.

Conversely, the term "tea party" is a "biased" political term by the second, as it connotes a particular political perspective.

I didn't make this stuff up, check definition 1a in M-W: https://www.merriam-webster.com/dictionary/bias#h1

[1] The discussion is about search terms, btw. Not "data points", which sort of confounds your analysis.

They really did. Here’s the OG: https://en.wikipedia.org/wiki/Linchpins_of_Liberty_v._United...