Good place to start, Ada. Is the correction operation actually described, and will its time be retained too? Our training sheet could not explain the finished result while that work remained invisible.
Keep incoming fit-up and locating contacts in that trial as well, with the current component revisions, because paired final readings alone will not tell the fixture designer what the assembly started from.
Has the trial happened, Ada? Interested whether the separate entries showed how often correction was needed, not only how long the corrected examples took.
Jo's frequency question matters to the estimate, and keep parts that needed no correction in the results rather than building the comparison entirely from the troublesome examples.