It depends on the sensitivity of the sensor, the thing measuring it, and its output characteristics, so it can be a complicated thing to answer.
That said, if you're measuring the sensor directly with the micro, it's probably not all that sensitive to a lot of noise (there's already a fair amount built in), so make sure your measuring instrument has a fairly high impedance and use shielded cables to connect the points and the device. Basically what you'd have to worry about is noise picked up on the measurement line, provided your source impedance is low enough that the added measurement of it doesn't draw the voltage down (and provided this is a voltage measurement), but unless your test points are in a shielded box, you're still gonna get at least a bit picked up on the board itself.
If it were me? Add test a test point for the signal and ground (a solder-in loop is great for oscilloscope hookups, but you can use wide through hole pads on the edge of a board or a bit of wire with a loop soldered to a blank pad), keep them fairly close together, and just clip the probe in normally. A scope isn't going to be super high resolution on the vertical scale (they're not designed for it), and your micro is probably reading its ADC with similar or slightly higher noise levels than the scope, so it's likely that your application is not sensitive enough for the probing to matter much. The easy sanity test would be to measure your device on the micro, then connect the scope and measure on the micro again. If the readings are consistent, your scope probe probably isn't injecting enough noise to make a problem.