Absence of evidence is not evidence of absence – and that affects what scientific journals choose to publish