A possibility here, which I learned in the 1970s from a Champion Plugs rep at the time: the inside of nearly all sparkplugs consists of a metal rod from the wiring end to the middle of the plug body where it and the the engine-side center electrode (often inconel) meet a 'plug' there, on the opposite ends. In non-resistor plugs there is a metal plug installed between them (good ones are pure copper, poor ones are other metals) and in resistor plugs this is a carbon slug of "R" ohms, not unlike a wooden pencil's lead in appearance. The connection is not a 1-piece thing: it is always 3 pieces so the plug body can receive differing parts for differing heat ranges, etc. in production without changing the machinery to make differing sparkplugs.
These inner connections are mere contacts, one pressed against the other, and in brand-new plugs they are often not a perfect contact (at first), The "plan" for this is: they will arc themselves into a welded unit in just a few seconds of run-time. The better plugs out there will do this at the factory (this used to be NGK, Motorcraft and Bosch) in a final-test station of a sparkplug-building machine (I used to build equipment like that in my day job).
Today, it seems that no one is testing ANYTHING, anymore. All this is a long-winded explanation that is pointed to this last statement, which I have experienced recently as accurate: "The sparkplug's internal circuit is not completed until the plug is installed and run for a few seconds." The impurities in the contact surfaces between the 3 components arc-weld themselves together with the first application of high voltage. Then the published specs come true, like magic.
I most recently found this in a set of new plugs for my Explorer: they all measured wildly different readings with a digital ohmmeter, with much closer readings using an analog meter (but still significantly differing), and after driving them a week, the 2 I re-pulled from the engine to test showed identical ohms. This sequence of results should remind one of a very important fact to keep in mind for digital ohmmeters: they provide almost no current to the DUT (Device Under Test) due to the mass-manufactured type of integrated circuit that is used inside of them (to make them cheaply): these chips use so little current to measure ohms readings that they will not penetrate any crud of any kind (even in brand-new switches, relays, and...sparkplugs), so the erroneous readings seen when using digital ohmmeters should always be taken with a grain of salt. If you have any access to a meter-type VOM, use it instead for these sorts of tests. This is one important reason why the "old" analog meters are still out there, and still very popular.
