We label epistemic status.
Belief, hypothesis, experiment and replicated finding are not interchangeable.
We preserve negative results.
A result shouldn't vanish because it became inconvenient.
We version material changes.
Research evolves; the record should show how.
We disclose methodological changes.
Changing an evaluation after seeing the result creates obvious problems. If a protocol changes, the change is visible.
We publish enough to be criticised.
Where privacy, licensing, security and contract allow, research includes enough for an informed reader to inspect what was done.
We distinguish internal from independent replication.
Running our own experiment again is useful. It is not independent validation.
We don't use conviction as evidence.
Even when we really like the philosophy.
We correct the record.
Visibly.