Data accessibility is not sufficient for making replication studies a matter of course