pporrasebi
commented
3 years ago Interaction confidence and parameter unit are treated as very different things in the CV. Both are children of the root term directly and operate in different ways. Confidence works as a set of categrized free-text annotations and the parameter units are fixed terms that need to be associated to a numeric field.
My suspicion is that confidence was set up this way because there is a huge range of ways people use to determine confidence, many of which do not use a quantitative output, but a qualitative one, so defining units for something that is going to need an annotation to describe how it was defined anyway was likely viewed as pointless. I pretty much agree with this and would advocate to leave confidence as an annotation. Also, I would heavily advise against using 'p-value' as a unit, which is not. p-values shoudl be reported with summary statistics about the test they refer to, not as a meaningless unit. I think p-values are over-interpreted and over-used in science in general, so I do not think we should contribute to the problem.
Regarding definitions, I have altered the definition of 'kelvin' to the latest SI redefinition, which is the following: "The kelvin (K) is the SI unit of thermodynamic temperature. It is defined by taking the fixed numerical value of the Boltzmann constant k to be 1.380649×10-23 when expressed in the unit J·K-1, which is equal to kg·m2·s-2·K-1, where the kilogram, metre and second are defined in terms of h, c and caesium frequency. It is equivalent to -273.15°C / -459.67°F."
How come there's not a single unit for confidence ? Can we please add, for starters, p-value ? or is there some other CV to use ? other than that, definition for parameter unit should not refer to kinetic constant units because, eg 'kelvin', (btw. any physicist reading kelvin definition will cringe :o/ I'd just write kelvin is SI unit of thermodynamic temperature, in order to to bypass intricacies and history of its definition) can be used in a context different than kinetic measurements; sadly, specific reference to kinetic constant makes underlying terms not applicable in other contexts :o/
it might also make sense to make, both, 'parameter unit' and 'confidence unit' children of 'unit'. Note, that 'confidence unit' should refer to generic measurement confidence, without referring to interaction and/or participant identification. In general, every cv term definition should be made as generic as possible.