The answers to the poll are not relevant, because it's impossible for anyone to accurately answer it.
1) No one knows exactly what's happening with the spider gene. We know that spiders and normals are produced from a spider X normal or spider X spider breeding. We know that no 'super' form has been produced--there are no homozygous spiders. We do not know why.
2) The percentage is the chance per egg that the animal will be something. It's not the percentage of the clutch that will be something. (For example, if you breed a pastel to a normal, each egg has a 50% chance of being a pastel--this doesn't mean half of the clutch will be pastels. It's only a chance. The first clutch I produced contained 13 eggs, and 11 of them hatched into pastels, with only 2 normals!)
The only way to solve this would be to get as many spider breeders as possible to start keeping statistics on their spider X spider clutches--detailed ones.
These should include the number of eggs in each clutch that go bad (if any), and the number of spiders versus normals that hatch. The larger the quantity of information we have, the closer we'll get to the true probability, which may reveal whether the super-spiders die in the egg, or are mysteriously absent with all breedings producing 50/50 chance per egg.
The problem is, people rarely breeder spider X spider. It's far from uncommon to lose an egg or two during incubation, so folks would tend to ignore it. Because it's a percentage chance per egg, of course some spider X spider clutches would have all good eggs, even if it IS homozygous lethal.