You could test this by setting up your panels in parallel but with a low-value resistor in each line, leading to a common load. You could measure the voltage across one resistor as that panel goes into shade, and see if it ever reverses polarity, which would indicate current is flowing back into the shaded panel. I think moffy is right that it would just contribute less and less current to the load. But I wonder what would happen if the maximum open circuit voltage of the shaded panel is lower than the voltage being produced by the full-sun panel. It seems that might be a problem, but I've never tested it.
I wonder if ideal diodes could be used in low-current solar systems in cases where a diode is actually needed, such as to prevent the battery feeding back into the panel at night.