The noise from the switcher isn't an issue for digital designs and easily dealt with for 10 bit A/D conversion with a 5 V reference because even the fastest A/D conversion I did averaged several measurements for accuracy and external noise rejection.
Not a switcher power supply as being discussed, but I once built a board with 12-bit
ADC and opAmps running off +/-6V linear v.regs [78L06,79L06], and with the negative
rail being supplied by a 7662 chip. In the prototype lashup, there was a lot of noise on
the ADC conversions, but that was solved by placing a ground plane under the 7662,
and also moving the 7662 further away from the opAmps.
So, having good ground planes and physical separation of the spike makers is 1000%
required.
In bread boarding I've found that those little breadboard jumpers are totally unsuitable for carrying any current above 100 mA or so, depending...
As they are 30 Ga stranded and with the combination of jumpers and breadboard connections typical are lossy.
I start with 5 V from an external PSU and after 6" of travel through jumpers and pins the voltage ends @ 4.5 to 4.75 V.
The Arduino female headers with the tuning fork inserts, combined with temporary
jumpers using circular endpins, makes for a perfect storm of iffy reliability. OTOH,
the usual 0.1" male headers, combined with suitable female pins, makes for a reliable
1 Amp [or 3 Amp] connection.