When the model in your head mirrors the observation. That’s truth.
That conclusion is unsupported. Only the facts are truth.
When your mental model matches your observations and makes accurate predictions about expected values for observations that indicates that your model is at least close to the truth.
However if you control the observations, you have to be careful not to introduce a bias. For example you might be tempted not to make certain observations that would falsify your model. Or you might be tempted to not make observations that would require you to expand your model. But in both cases you admit your model doesn’t match reality and what good is it at that point?