It does in cases such as this. The goings-on immediately surrounding the MAX232 (the original, and not the later "A" version and beyond) really makes a lot of noise. Two things are to be considered. One is that the MAX232 switches at a frequency far from the band in which the MCU makes its noise, and the other is that it is quite a bit noisier than the MCU, in terms of current. This shows up if one puts a low-value resistor in the Vcc path. In the case of the MAX232, it's helpful to have that same effect at Gnd, as it motivates/enables the Vcc-to-Gnd cap to do its work within the MAX232's domain yet not impact the current on Vcc so much. That causes the noise current to circulate through the cap instead of through the entire power net. Normally, it would cause offset spikes in the GND line that would make Gnd a really nebulous concept, but in this case, with a resistance half of what it is in the Vcc line, or less, it keeps the current noise in that loop. Signal in or out doesn't seem to be degraded in any way.

I'd suggest you try it if you have nothing better to do. I was amazed!

The need for such precautions goes away with the MAX232A and later. I guess the currents being switched are much lower.

RE