Keep in mind, that meter is also measuring the resistance of the wires and the resistance of the contacts to the battery. It's not unusual for contact and wire resistance to be equal or greater than that of the battery. To get a true reading, you need to to calibrate with respect to contact and wire resistance. That may be the case for the measurement in that photo, but there's no way to tell just by looking.
I would probably trust the specs listed in a data sheet before that of a hobby meter with a regular battery holder. Even so, it's not that far off. The reading indicates 18m Ohms per battery which is pretty close to the impedance spec stated in the data sheet.
Also, the specs in the data sheet are internal impedance measured with a 1 kHz frequency generator and analyser. This is not the same as the voltage drop you would see using that number as a resistance value. Resistance and impedance are not the same thing, though they can be under purely DC conditions.
The internal impedance spec is reflective of a battery's DC performance and it's a good number to use for comparison purposes, but it's not necessarily going to provide the actual voltage drop seen when the rubber hits the road. In some data sheets you see the DC resistance value as well as the impedance value. DC resistance is usually a bit higher. That makes the number seen in the photo quite reasonable if that meter is measuring DC resistance and not impedance as stated in the data sheet.
As far as "cherry picking", there may be some of that going on. They usually spec the prototypes and then put limitations on how much units can vary from that coming off the line. There's always a number of cells coming off the line that end up in the recycle bin so you can end up with one either better or worse than the stated specs. A quality maker puts a pretty tight tolerance on that kind of thing. This is the primary difference between a Japanese and Chinese cell, the consistency in quality.
In my own experience with quality cells, I've not found the impedance spec to vary much from one cell to the next. I've have yet to end up with one that has a markedly higher internal resistance than another. It's something I pay pretty close attention to in testing.