Ideally, your data comes from multiple labs, replicating the same research hypothesis.
Consider these possible scenarios: It is easier to get funding for new research, than to replicate and confirm someone else's research. A group of people in the same lab will be more apt to support published research of a co-worker, than look for holes and problems in their methodology. The more publications that have been published based on a theory - the harder it is to get alternative theories published. People with large reputations are less apt to be questioned, or to tolerate dissent. Do you think such scenarios might distort or bias what data is collected, how it is collected, how it is reported, and whether it is published?