popular opinion as evidence

[Coffee Jesus]

For a while, popular opinion was that the world was flat.

In the Middle Ages, people believed in Spontanious Generation. Rats spontaneously generated from piles of rags, maggots from meat.

People still believe hair and fingernails continue to grow after death, old windowpanes appear wavy because glass stays "liquid" and flows downward very slowly, taste zones on the tongue, that Columbus was concerned about falling off the edge of the earth, that a Brontosaurus was a type of dinosaur... The list of "facts" goes on.

The popularity of an opinion has no bearing on the truth of that opinion.

Confirmation bias. We never stop to think about the commonly accepted claims that were never overturned. They are the water we swim in.

Any random claim you could have pulled out your ass will probably be false. Suppose its probability, P(T), was only 0.01 to begin with. If P(O) · 3 = P(O|T), then its probability has only increased to .03 per Bayes' formula. Its probability is still very small, but it did increase.
P(T|O) = P(O|T) · P(T) / P(O)
P(T|O) = .75 · .01 / .25
P(T|O) = .03

Furthermore, the degree to which the claim's probability increases depends on the relationship between the event of the claim being true and the event of people supporting the claim. If some janitors tell me a clown ran down the hallway, the probability of that increases by a lot. Not so much if the janitors tell me there are exactly three universes.

There is a difference between popular opinion, and consensus based on observation, testing and falsification.

The "popular opinion" of religion and god claims is merely mob rule by placebo.

The consensus of science isn't dependent on popularity.

There is a HUGE difference between when a scientist says "Most scientists accept"

And, "Most people believe".

To attempt to equate the two is absurd.

I didn't equate the two. If I know that it became popular through reliance on observation, then the probability increase is larger than when I don't know why it became popular.
But the probability increase isn't entirely negated, unless you can show that I have no reason to think any of its popularity is due to a reliance on observation. You would do this by showing that we could expect some degree of popular support even in the false condition, even if you leave the expected amount blank. Then I would have to show that the expected amount is still below what is actually observed, even after taking your explanation into account.