It depends on how the replication is done, but the big replication projects typically use very large sample sizes compared to the original studies, so their err

It depends on how the replication is done, but the big replication projects typically use very large sample sizes compared to the original studies, so their error bars are much smaller.