Array 8x8 pixels

relative to S-2 images, I would ask if is possible to “quantify” the “dominance effect” of one pixel on another if we considered ONE pixel. This effect produces a not true value of DN (one for each spectral band). This effect, I have read elsewhere, can be minimized if we considered an array of (minimun) 8x8 pixels.
It’s true? There is some official documents of ESA attesting this?
Thank you.