At the dawn of the modern age of science, a few hundred years ago, accounting for the motion of the planets was a mystery, but one driven by a flawed theory. It was thought, going back to the ancient Greeks and Plato, that the motions of the planets, being otherworldly and celestial objects, must be perfect and therefore circular. Unfortunately, actual observations were hard to reconcile with this notion. The ancient astronomers could have fudged the data to make it conform to the theory, but that would have been unscientific, so they fine-tuned the theory to try to make a better fit. Almost two millennia ago, Ptolemy refined the concept of circles within circles, or epicycles, to try to develop a model that would explain the observed planetary motions. The theory reached its height half a millennium ago when Copernicus, with the insight that the earth orbited the sun, like the other planets, came close to modeling planetary motion by adding new epicycles, albeit with a different model for each planet. But it was a very complex system, and still wasn’t quite close enough.
Kepler resolved the issue by demonstrating that the best fit of the motion was not circles within circles, but rather simple ellipses. He came up with simple but powerful and explanatory laws that described the motion of the planets as a function of their distance from the sun. Newton in turn used this finding to validate his own universal theory of gravitation.
But it still wasn’t quite good enough. For centuries, the innermost planet, Mercury, stubbornly refused to conform perfectly to Newton’s laws, and many more modern astronomers postulated a hidden planet elsewhere in the solar system that might account for the discrepancies; they didn’t abandon Newton’s theory. However, despite years of trying, they could never determine its location or mass. But despite this frustration, they never yielded to the temptation of simply denying the planet’s mercurial behavior — they continued to refine the theory, no matter how difficult.
About a century ago, another physicist, Albert Einstein, came up with a new theory of gravitation. A key part of it is that Newton’s laws must be adjusted slightly to account for the near presence of large masses. By Einstein’s new theory of general relativity, of which Newton’s earlier theory was simply a special case for velocities much less than that of light and locations not adjacent to very large masses, Mercury’s motion was perfectly explained by its close proximity to the sun.
Over thousands of years, at each step, the response of the scientists was to continually adjust and refine their theories to conform to the data, not the other way around. This is how science is done and how we developed the knowledge that has given us such tremendous and accelerating scientific and technological breakthroughs in the past century. It is occasionally reasonable to throw out a bad data point if it is in defiance of an otherwise satisfactory model fit, as long as everyone knows that you’ve done so and the rationale, but a deliberate and unrevealed fudging of results in an attempt to make the real world fit one’s preconceptions is beyond the scientific pale. Journal articles have been thrown out for it; PhD candidates have lost their degrees for it.