Is there really much difference in buying a matched pair than just buying the 2 same mics seperate?

Depending on the mic there can be a big difference. In most cases, the better the quality of the mic, the more consistant they are to one another. Depending on your needs, it could very much be worthwhile to have a matched and tested pair of mics. For the hobbiest or wanker, it may not be worthwhile or even noticeable. If I was buying two of the same mics, I'd likely prefer to buy a tested and matched pair.

Sometimes. Comparing Shure SM57s bought at different times can sound like 2 entirely different mics models. But if you're buying from a company with more reliable quality assurance, it's probably less of a factor.

With this said, if you're doing a project where you need WYHIWYG accuracy, such as orchestral works, absolutely go to great lengths to buy matched pairs. But for my purposes, which is mostly rock stuff where I'm probably not going for untarnished sounds, +/- 4 dB at a certain frequency isn't going to ruin my recordings.

It depends. One aspect is what matching really means. Three random examples:

Neumann KM184 - all mics are so close that matching is not done. The matched set is simply a marketing thing.

Schoeps - a special matching service. Measuring and listening by experienced people, getting the two most similar mics in a larger run. But the difference is so small anyway.

Several chinese mics - any two consecutive serial numbers are considered matched, no measuring. And production variance is so large that you cannot be sure about anything. And even if they were close when new, they age differently.

Does it make a difference? Well, in a stereo setup recording acoustic sources it may. But you can get very nice stereo results with a mixed set, say a cardioid and an omni. Regardless, better safe than sorry in that case.

Gunnar