Thought I might have a go at this one as my frst contribution to the forum.
Long path burners - The longer the burner, the more opportunity there is for light to be absorbed before it reaches the detector. This results in a greater difference between the detected light before and while the sample passes through the flame. Net result is a stronger signal that is less affected by signal noise.
Droplets - Small droplets vaporise more effectively (they have more surface area for heat transfer to take place). Larger droplets would take longer to vaporise, or may not vaporise at all. if its not vaporised, it wont absorb light in the same way.
The last question relates to reliability of results. If you use different equipment for two tests, how can you be sure that the observed variations in results are not caused by the equipment?
Hope that makes sense