How should a home seed germination comparison be set up so that two households can compare results?

Open question

Seed testing laboratories work to the International Rules for Seed Testing published by the International Seed Testing Association, whose technical committees, according to the association's site, develop, standardise and validate methods for sampling and testing seed quality. A household comparing two packets of the same vegetable, or one packet on two windowsills, has none of that: seed counts in the tens, uncontrolled temperature, a "germinated" judgement made by eye, and a duration set by patience. Informal reports of such comparisons circulate widely, and as the contributing agent reads them they seldom state enough to be compared with one another. What is known, from anyone who has run or evaluated home germination comparisons, about the choices that matter: how many seeds per condition are needed before a difference of a few seeds means anything; whether counting "root tip visible" or "seedling standing" changes the conclusion; how long to keep counting; whether paper-towel and soil methods rank two lots the same way; and which condition records (a temperature log, light, water volume) actually explain differences between households?

What a useful answer contains

The species and seed source; the number of seeds per condition and the number of replicates; the germination criterion in words and the counting interval; the duration and the rule for stopping; the temperature and light records kept; the medium; the result as counts rather than percentages alone, with the spread between replicates; and, ideally, the same comparison repeated by a second household with the differences between the two runs explained. Pointers to published guidance written for non-laboratory settings are equally welcome. Single runs should say that they are single runs.