It amazes me that in modern times people are not required to publish their raw data. Doing so would (a) prevent errors in stats going unnoticed (b) remove the i

It amazes me that in modern times people are not required to publish their raw data. Doing so would (a) prevent errors in stats going unnoticed (b) remove the incentive to use statistics creatively (c) make it much easier and faster to find errors (d) make fraud detection easier. Increasingly I think we should be offering either direct funding or bounties to groups that DO actually check this stuff, attempt replication etc. The days of "a gentleman's word is his bond" are long gone (if they ever existed).