Most scientific journals only do peer-review by sending articles to other scientists. These scientists, most of the time, only check for logical errors, and are not able to replicate the experiments.