Consider for example the case where the ISM of a galaxy contains a lot of dense, cold gas. This may lead to strong star formation in these clouds, and the resulting massive stars will heat-up the dense gas, both with their energetic stellar winds, their radiation, and their SNe explosions. Alternatively, if a galaxy has no strong star formation, the hot gas will eventually cool and form denser clouds. So you see that star formation may act as a (negative) feedback process, regulating the amount of cold gas and the rate of star formation in the disk.
This intricate process is not well understood, but let me wave my hands a bit and make the following argument. The energy of a SN depends on the type of star, not the type of galaxy. Hence you can imagine that SNe explosions (for example) have more effect on the gas distribution in a small galaxy, with little mass (mostly in dark matter) to constrain hot gas in it, than in a very massive galaxy. If there are sufficient numbers of recent SNe in a galaxy with not too deep a potential well, then such a galaxy may have trouble holding-on to the hot gas generated by the SNe, and a galactic wind may develop.
And so the net effect of feedback in a small galaxy might be that a significant fraction of the gas is removed from its potential well, leaving mostly dark matter and some residual gas, and the rest of the stars. Indeed, we find observationally that such dwarf galaxies contain very little gas or stars, in comparison to the amount of dark matter in them. Feedback from SNe is thought to be the culprit.
In more massive galaxies, the potential well is too deep for a galactic wind to remove a lot of gas, and so more massive galaxies tend to hold-on to most of the heated gas. In a galaxy like the MW, the effect of SNe is to generate a galactic fountain. Gas heated by the combined effect of several SNe rises perpendicular to the disk, and starts to cool as it expands. When it has cooled sufficiently, it rains back onto the disk. Crucially: it eventually comes back to the disk.
As I explained, the SNe not only heat gas, they also produce metals. If the hot gas can escape the halo of a galaxy, then presumably so can the metals. And so small galaxies cannot hold-on to the metals they produce, but more massive galaxies can. At least, that is what we expect. But indeed, it turns-out that the more massive galaxies tend to have higher metallicities than the lower mass ones.
Let me make clear that this certainly is not the whole story. Magnetic fields and energetic cosmic ray particles4.9 contribute significantly to the energy density in the plane of the MW. So they are likely to be important in understanding the dynamics of the ISM, yet they certainly make an already complex and non-linear system even more intractable. And hence are usually ignored, may be to our peril.