I'm not sure I understand your contention.
Including an AMFA worked month is entirely the purpose of the comparison. This is a one month YOY comparison, done so to equalize any seasonality and operational impacts that may skew comparitive results. By comparing delay rates, (rather than absolute delay totals), the effect of volume is also mitigated. The result is a true comparison of performance between the two mechanic workforces as it relates to mechanical delays.
To repeat, March 2005 figures were with AMFA mechanics fully employed, and March 2006 figures were with replacements.
Delays are assigned a delay code at the time of occurrence. As I stated, one could claim that coding criteria has changed to make the YOY comparison look better than it actually is. But, if that's the case, then why wouldn't NWA be touting these figures in press releases to counteract the AMFA drumbeat of an "unsafe and unreliable" airline? Also, what operating group is going to willingly be burdened with the delays that are coded away from Tech Ops? Every delay has to go somewhere, since the sum total of all of the system delays need to be accounted for via a delay code. Every operating group leadership team has to meet a targeted delay rate as part of their departmental goals, and no group would accept additional delays that don't belong to them just to make some internal(and unpublished) figures look good for Tech Ops.