I can't imagine myself ever conforming until it was less than 1/8 as blatant as the example image.  Assigning a >50% probability to the majority being correct seems way too generous, because I have no strong evidence that they're not lying, and a high prior on my ability to see linear distances on a 2D page.

Did the 100+ replications collect any data on what sort of people are more conformist than others, besides the gender gap?