It is fairly well-described in Naval Aviation Combat Statistics World War Two. Claims for aircraft destroyed in the air quoted below:
[Airborne enemy aircraft claimd destroyed by naval aircraft, in aerial combat only. Planes destroyed by own anti-aircraft fire or in suicide crashes are not included. Enemy aircraft reported as "probably destroyed" are not included. Squadron claims, as made in ACA-1 or other action reports, are the basis for these figures. They thus represent the evaluations only of the squadron intelligence officer, squadron commander and, in some cases the air group commander. However, rarely was there any further evaluation by higher authority of squadron claims with respect to airborne enemy aircraft.
In evaluating pilot claims for ACA-1 reports squadron intelligence officers were instructed to follow the definitions of "destroyed" established for the command or theater. Subsequent to early 1944 this was the standard Army-Navy definition that the plane must be seen to crash, disintegrate in the air, be enveloped in flames, descend on friendly territory, or that its pilot and entire crew be seen to bail out. Prior to this time the definitions varied between commands, but the definitions used in the principal naval theater (SoPac) were at least equally stringent. The degree to which squadron intelligence officers and commanders succeeded in eliminating duplicating and optimistic pilot claims is not known, but it is believed the amount of overstatement is relatively low. Since 93% of all enemy aircraft claimed destroyed by Naval aircraft were claimed by single-seat fighters and the bulk of the remainder were claimed by two-place dive bombers and by lone search planes, the tremendous duplication of gunnere~ claims experienced by air forces operating large formations of heavy bombers with multiple gun positions is largely eliminated. Duplication of claims between fighter planes can be more easily controlled by careful interrogation.
Over-optimism has always been difficult to control. During the early part of the war, before standard definitions were in force, before full-time trained Air Intelligence Officers were available to apply them, and before the need for conservative operational intelligence was fully appreciated, action reports may often have overstated enemy losses. Evidence from the Japanese has tended to indicate that in some of the early actions, and even as late as the Rabaul raids of early 1944, there was such overstatement.
It must be remembered, however, that the bulk of Naval aerial engagements in the Pacific did not involve the mass combat of Europe. Even the large-size engagements seldom involved more than 30 of our planes against 30 of the enemy's at any one time within visible range of any one point. By far the greatest number of engagements involved only 1 to 8 of our planes, or the same number of the enemy's. Thus, in the main, the claims under this heading, offset as they are by the exclusion of planes classified as "probably destroyed", are believed to be near the truth, with only local exception, and to be as conservative as those of any major air force.]