In section 2 of Douglas-2007-rejecting-value-free-science, she emphasizes that in doing science, scientists must make choices at every stage of the process.
- choosing methods = which experimental design? which measurement technique?
- characterizing data = what is being targeted? other factors to include?
- interpreting results = what does the data mean? what background assumptions are considered?
Douglas notes that scientific papers are structured to reflect these stages (methodology, data, results) but scientists rarely make explicit the choices they make. Papers present the chosen path as if it was the only option, rarely mentioning the alternative paths that they might have taken.
Why do scientists avoid discussing choices =
“To discuss the choices that they make would require some justification for those choices, and this is territory the scientist would prefer to avoid … Because scientists do not recognize a legitimate role for values in science (it would damage “objectivity”), scientists avoid discussion of the choices that they make.” p. 6
Choices involve non-epistemic value judgements
Douglas’ main argument is that the choices made in the internal stages require the consideration of both epistemic and non-epistemic values. Choices involve risk of different kinds of errors, and if one is to weigh which errors are more serious, one will need to assign values to the various likely consequences. Thus, values become an important factor in the making of internal scientific choices.
In general, in contexts with recognized uncertainty and significant chance of error, we hold people responsible for considering the consequences when making decisions.
- In the emergency room, hospitals tend to avoid as many false negatives as possible when it comes to potential heart attack victims.
- the justice system attempts to avoid false positives when it comes to convicting people, because wrongful conviction is considered worse than wrongful acquittal Different contexts weigh errors differently based on value judgements about consequences.
Similarly, Douglas argues that we should hold scientists to the same standards of everyone else, and thus scientists should think about the potential consequences of error
“if we want to hold scientists to the same responsibilities the rest of us have, the judgments needed to do science cannot escape the consideration of potential consequences, both intended and unintended, both epistemically relevant and socially relevant.” p. 10
Rat Liver Tissue Example
Douglas illustrates the consequence of errors from choices with the following case:
Rats were exposed to a chemical dioxin and after two years of dosing, the rats were killed and autopsied. The liver tissue slides were evaluated for cancerous lesions by three different groups (over a period of 14 years) and produced different results because they applied different standards for what counts as a cancerous lesion
The choice being made = how strict should the criteria be for classifying a lesion as cancerous?
- avoiding false positives = risk of false negatives (missing real cancer) and the dioxin may appear as less potent as a carcinogenic, resulting in weaker regulations
- avoiding false negatives = risk of false positives (classifying non-cancers as cancer) and the dioxin may appear as more potent and dangerous, resulting in stricter (potentially more burdensome) regulations
This demonstrates that there is no neutral way of characterizing the lesions and that even in describing what appears to be “raw data”, scientists are actually making choices weighing consequences that involve non-epistemic value judgements such as
- how bad would it be to falsely accept as cancer
- how bad would it be to falsely reject as cancer
Value Evidence
Douglas emphasizes an important distinction that values and evidence are not the same thing.
- Values = statement of norms, goals, desires (what ought to be / what matters)
- Evidence = descriptive statements about the world
But value judgements are needed to determine whether a descriptive label (evidence) is accurate enough given the uncertainty and stakes involved