I suspect Classicist may have put his finger on something here.
I am an accountant, and a number of years ago I did was working on a business plan for a "direct marketing" organization. They wanted to know statistics about turnover for sales organizations. So I began a search into organizations like Amway and Shaklee, etc. What blew me away, was that they all basically experienced 100% annual turnovers. In other words, suppose one had a sales organization of 1,000,000 strong worldwide, and they still had 1,000,000 after one year. Then you look at the number of new recruits they took in during the one year period. Let's say it was 1,000,000. That would give you the impression they would have 2,000,000 after one year. Yet the statistics would show they only had 1,000,000 after one year. What this means or implies is that they would have lost 100% of their sales force, but would have had to recruit a new batch of 1,000,000 to replace the ones they lost over the past year, just to keep a level population.This is called "attrition" or "turnover".
I have not studied the JW statistics in many years, but I would bet there has been a high turnover rate in their numbers over the years. People who are recent converts to the JW faith would tend to be more active in the publishing work than those who have been doing it for a while, for the simple fact of their fresh conviction and enthusiasm. After a while, some of that gung ho wears off. Meanwhile others leave, or are disfellowshipped. But they are replaced with the latest crop of converts.
Now in the context of a year, those who reported field service at the beginning of the year would be counted. Then, throughout the year,some would quit, some would die, some would become inactive, while others got disfellowshipped. But to replace them would be the new converts. The question I would have, therefore, is if their time is counted, and each one is statistically counted as 1 person, but the other ones who were there before also reported time and are counted as 1 person (but later in the year quit, etc.), would that not inflate their numbers in terms of the number of publishers? I don't know the methodology they actually use these days, but I do have my suspicions.
(Actually, I was under the impression they counted their "members" in terms of the number of baptisms, but cannot confirm this for sure.)
Rod P.