My intuition would be to test the sample that wikipedia editors do reference rather than the sample they might reference - but I think either methodology is an

My intuition would be to test the sample that wikipedia editors do reference rather than the sample they might reference - but I think either methodology is an improvement over "argue about it and then summarize the argument".