Each journal has a set of permanent reviewers which may review any paper they think themselves competent to review. External reviewers can also be recruited, either by authors or by the internal reviewers. External reviewers should fulfill the following criteria:
1. They should have sufficient statistical competence to evaluate the paper. This doesn't mean that they must understand every method used in details, but they should understand most and have some idea about the rest.
2. They should have expertise in the specific topic that the paper covers. This should be interpreted broadly.
3. They should not be recent co-authors of any of the authors. The interpretation of recent is up to the review team to decide.
If an external reviewer has been recruited, the internal review team may veto the person to avoid authors recruiting obvious partisans or incompetent people. The internal review team should follow the guidelines above in their decision making.
During the early stages of the journals, it was necessary to ignore (3) in order to have sufficient reviewers. As the size of the review teams grow, it should become unnecessary to ignore (3).