Here is a quotation from Harvey's Techniques of Trout Fishing and Fly Tying, revised edition, pages 20-21:
"I made my way to the head of a flat pool I knew held a good population of trout. I sat down upon the bank and started a chum line of [Japanese] beetles. It wasn't long until I had about a dozen trout taking every beetle that floated down. Now I took twenty-inch pieces of hard nylon of all sizes I had with me and inserted the end of the nylon into some of the beetles and sent them down the chum line. Since there was no drag on the nylon, the browns picked off most all of the beetles that were attached to the nylon.
This was not my idea. I had read of a similar experiment conducted by John Crowe and I wanted to check it for myself. Let me tell you, I was really impressed. Most of what we read today and what most fly fishermen believe is that the fine terminal part of the leader is more invisible and that this is the reason for using fine tippets. In fact many articles have been written that recommend that the size of the terminal tippet should be determined by the size of the fly one is using. This experiment blasts that theory because the stream-bred browns did not refuse a live beetle on the free-floating and drag-free pieces of nylon is sizes up to .015."
A shorter description of this experiment can be found in _George
Harvey: Memories, Patterns ands Tactics_, as told to Daniel L.
Shields, p. 71.
In rare circumstances a leader can spook fish by casting a shadow and reflections (from the depressed meniscus of a floating leader) on the stream bottom in very clear and shallow water. But this is a very specific situation that does not apply in 99%+ of fisheries.
In regards to the length of the leader, you can lengthen the leader using the same length of the same tippet without lengthening the tippet itself. You do this by lengthening the butt section. You will find that this rarely reduces the rate of fly refusal. In fact, using a shorter overall leader but a more flexible leader design of a furled leader will suffice. A proficient caster does not land the fly line in the window of the trout.
Finally, the proof of the pudding as far as micro drag is concerned is to NOT change any part of the leader or tippet but to change the cast itself. A parachute cast/mend done from upstream of the fish will catch the is demonstrating that micro drag was the cause.
The downstream parachute cast/mend is the solution to this problem.
Position yourself up and slightly across from the feeding fish. Then make a parachute cast down and across and the feeding lane. The fly lands upstream of the fish. Skate the fly into the proper feeding lane, and then drop the rod tip to "feed" the fly to the fish in rhythm with his rise. Trying to remove all drag with a slack line, being highly accurate, and timing the fly for a rhythmical feeder is nearly impossible when casting from below the fish.