Probably, what they are reading is some article in popular press about some data which is not totally explained by General Relativity - sort of information requiring dark energy/dark matter etc.
In the following I want to address the way a scientific theory becomes accepted as a good description of nature but only until some new data contradicts the predictions of the theory in a fundamental way. Then scientists try to formulate a new theory, either extension of the old theory or something radically different, which has to pass the usual tests - this is what I want to discuss in the following. I have reproduced some slides from my talks on Einstein and The Theory of Realtivity:
This is a painting (refer to Wiki) showing the general belief during the 19th century that horses legs are splayed when they gallop. New evidence from Muybridge in 1887 shows that the legs are actually gathered together and are not splayed.
The usual process of constructing a scientific theory is best illustrated through a flow chart - a series of steps:
1. Define/Identify the problem
2. Form a hypothesis/Theory
3. Make observations to test the predictions of the Theory
OR Conduct some experiments to test the hypothesis
4. Data collected from 3. is analysed and results obtained
5. Do experiments and data support the hypothesis/theory?
5A: IF YES Then the theory is good and you can publish the results
5B: IF NO Then you must first check if the experiment has any faults which might be affecting the results. If the experiment is faulty then correct it. If experiment is good then
6. Discard the Theory and start again by formulating a new Theory (Go to Step 2)