First, it's important to realize that all proposed changes to physics need to be tested, whether they amount to adding new stuff to the universe or modifying equations that have worked fine thus far. Suppose someone comes along and says, "I can explain this supposed dark matter by modifying gravity," and lots of theorists agree. Great. Now observers will go out and make observations that distinguish between the old theory and the new theory. Remember, observational astronomy : theoretical astrophysics :: experimental physics : theoretical physics. If nature didn't push us in the direction of looking for dark matter particles in lab experiments, we'd be spending more money building telescopes or doing other things to test gravity. If there is a finish line past which we cannot think of any science to do, it certainly isn't in sight.
rather than modifying our basic equations in physics for the macro-scale
There are people who think dark matter is just a modification of how gravity works. The primary such theory is MOND. And indeed tests of such theories are proposed. Alternate theories such as MOND have been steadily falling out of favor, however, due to the overwhelming evidence in favor of a new matter component.
We spend large amounts of money on the search for Dark Matter and so far haven't found a single thread of evidence it's real
There is an overwhelming amount of evidence pointing toward some missing matter. The Wikipedia article nicely lists the main observations contributing to this conclusion, going all the way back to the early 1930s:
- Originally astronomers noticed the motions of galaxies in clusters, assuming the systems were in virial equilibrium, were too fast unless there was unaccounted-for gravitational mass.
- Around the same time, a similar discovery was made for stars orbiting galaxies.
- Gravitational lensing measurements confirm these masses, even if you thought the clusters weren't in equilibrium.
- In one striking case, the gravitational mass is clearly offset from the visible mass, and models that try to adjust the large-scale behavior of gravity cannot simply displace the center of gravity from the center of mass.
- The large-scale structure of the universe -- as seen in the distribution of galaxies and intergalactic material -- looks as we would expect from having a large amount of non-interacting mass.
- In fact, the evolution of the universe as a whole only makes sense with a non-interacting matter content, as otherwise the cosmic microwave background and supernova distance measurements would be very different from what we observe.
All this evidence has been checked over and over again, and most agree that it all checks out. In fact, all these different lines of reasoning lead to the same estimate for how much dark matter there is. Also note that scientists didn't just jump on the bandwagon 80 years ago -- it took decades of collecting evidence for the majority to become convinced that there really is some form of dark matter.
At what point do researchers in physics make the leap from wild theoretical ideas to physical experiments?
This is a valid question. The rough answer is when the theory is important enough that verification or falsification is worth the resources required to test it, taking into consideration what else can be accomplished with those same resources. If no one to speak of believes the theory, testing it is a low priority. If everyone is so convinced of the theory they would doubt the experiment before they doubted the theory, then testing is also a low priority. If there are simply more interesting things (or better government-funded things) for experimentalists and observers to do with their time, they probably will spend less time on the theory in question.
It's also worth highlighting that a theory must be testable in order to be tested. The common gripe from experimentalists about string theory is that it does not lend itself to many feasible experiments. Direct detection of dark matter, on the other hand, is not beyond the realm of possibility: There exist masses and cross sections for hypothetical weakly interacting particles such that
- such particles would explain indirect observations to date,
- such particles would not have been directly observed with any previous experiments, and
- such particles would be detected with experiments we can build today.
Such experiments should either verify the whole theory or falsify part of it. Whatever the outcome, they are intended to produce results.