The JRD thesis applies as much to observation and evidence as it does to the conclusions of inference.
Many arguments for values playing a role in science depend on uncertainty or inferential risk. These arguments have been criticized for treating values as subordinate or secondary to evidence. One version of the argument introduces values as a tie-breaker, and the criticism most readily applies to this version. Another, stronger version of the argument appeals to our duties to have true beliefs and to avoid false beliefs. Balancing these duties inevitably involves value judgement. Standard formulations of this argument only apply to conclusions drawn from evidence, so the criticism applies here too. I argue, however, that the argument holds as much for observations and evidence as it does for the conclusions of inference. So the argument does not ultimately give an illegitimate priority to evidence.
The first on-line draft of this paper was posted 3dec2016.