The article is a good illustration of one of my gripes about cultural group selection. Proponents often muddle together genetic and cultural groups - along the lines of the fallacy of the extended genotype.
Human DNA is influenced by kin selection. Similarly, memes which are influenced by cultural kin selection. Some like to describe kin selection in terms of group selection.
However, to apply cultural group selection to groups of humans - it's a conceptual muddling together of the cultural and organic realms. It is usually best to keep these realms separate - applying organic group selection to genes in the organic realm and cultural group selection to memes in the cultural realm.
This is not to say that the two systems don't interact, rather that attempting to mix separate systems - with independent gene pools - when discussing kin or group selection doesn't make much sense - and just leads to muddle and confusion.
Compare with smallpox. Guns wiped out many native Americans, and smallpox germs wiped out many more. If the gun memes are adaptive on the level of "groups", should we similarly say that smallpox is adaptive on the level of "groups"? The conventional response is to say that this is a wrong question. If we look at the features of the smallpox virus, they are adaptive mainly at the level of the smallpox virus itself. The fact that the virus acts to wipe out groups of humans is an irrelevant epiphenomenon - as far as the adaptive features of the smallpox virus goes.
It's best to look at culture the same way. Culture is adaptive for itself first and foremost.
Now, it is possible to have adaptations in one species that benefit another one. We see this in domestic crops, for instance. Maize's strange cobs are there to benefit the maize plant - but there's also a sense in which they are there to benefit humans. It is a legitimate question whether Maize cobs benefit humans or human groups - and George Price's methodology for approaching this question could be applied if you had a specific breakdown of humans into groups in mind. But surely it is pretty dubious to say that maize cobs are there primarily to benefit humans. Maize cobs are primarily adaptive to maize genes. They've just found a way of manipulating humans into planting maize seeds.
Many memes have a much more negative effect on the fertility of their hosts than maize cobs do. If you think meme adaptations are primarily there for humans then you lose the ability to explain phenomena such as the inverted J-shaped curve of meme adoption and the demographic transition in Japan. Memes do not exist primarily for the benefit of humans - or for the benefit of human groups. They exist for their own sake.