I was listening to a podcast that made a really good point about the varying reliability of different sciences. In psychology, for example, a theory might have

I was listening to a podcast that made a really good point about the varying reliability of different sciences. In psychology, for example, a theory might have been validated in only a small experiment and fails when replicated. Meanwhile, something like Maxwell's Laws of electromagnetism are replicated trillions of times a day. Every time your cell phone or wifi works correctly, that is another replication indicating that Maxwell's equations were correct. When the science becomes engineering and we can build functioning things based on the theories, that is a good indication that they are correct. Maybe you didn't personally verify that the Earth is round, but the people who built the GPS system used that fact. If the Earth were flat, the GPS wouldn't work. So far, no flat-earthers have built a functioning GPS based on the Earth being flat.