The book is internally inconsistent, with the author cherry-picking "facts" to support his own prejudices. In some cases the same data is used for mutually exclusive conclusions in different chapters. Basically, he making up conclusions and then throwing any shit against the wall in the hope it sticks, with no regard for what he's already said in the same book.
It's pseudo-science, designed to let complacent readers validate their prejudices, and pat themselves on the back.
Best obtained via LMDE