At one time I had over 30 years' worth of dyno run documentation, some of it recorded by my father, and some I did myself, from the high school auto mechanics shop where we both taught from 1971 to 2010, with a 7-year gap between us. That shop was equipped with a chassis dyno, a water brake bell housing dyno, and all the equipment needed to build high performance engines except a crankshaft grinder. During my tenure, an over-zealous assistant principal discarded a filing cabinet full of "obsolete" paperwork that included dyno run documentation on at least 100 engines! He didn't get a Christmas gift from me again!
Those runs proved one thing beyond the shadow of a doubt- - - - -headers, compared to stock cast iron exhaust manifolds are good for three things- - - -increasing the under hood temperature, blowing out gaskets, and burning up spark plug wires. They also rust out in a hurry unless they're stainless steel. The only measurable power gain is achieved at RPM's far above anything the average street engine ever sees, and to get a power gain that's worth the effort, each header tube must be exactly the same length from the exhaust port to the collector. The best way to achieve that is to use a collector about 6" in diameter, and let the pipes from the last three cylinders extend into it. I ran a 460 cubic inch "440" MOPAR pull truck engine with that setup, and sized the diameter and length of all the header tubes to hold exactly one cylinder's volume of exhaust gas. The tubes were about 6 feet long each. We won several pulls in the Super Stock class with that engine- - - -running one carburetor and gasoline, and producing just over 700 HP at 6500 RPM.
BTW- - - - -that "one carburetor" which was required by the pulling class we were running was not a big Holley, Edelbrock, or Carter- - - -it was a Qjet for a 455 Buick with a little bit of jetting change to accommodate the "Street Tunnel Ram" intake, radical roller cam, and ported heads were running.