Talk:Result range
Should inclusion or exclusion be the default?
| Fact discussion: . |
|---|
| Opening statement: Inclusion of all values should be the default
Closing statement: Accepted. (A closing statement, when resolved, should be updated to the main page.) |
| Argumentation:
←--1: . Theoretically, when we start to describe a variable, we know nothing about its result. Therefore, it is logical to assume that it may be any value. Information is needed to exclude values from the domain. --Jouni 21:55, 28 September 2008 (EEST) (type: truth; paradigms: science: defence) ←--2: . Exclusion is analogous to falsification, which has been shown to be an efficient way to operationalise the accumulation of scientific information. --Jouni 21:55, 28 September 2008 (EEST) (type: truth; paradigms: science: defence) ⇤--3: . If the burden of proof is put to the person who wants to include values, it is easier to operationalise the problematic concept of 'impossibility'. See the following discussion --Jouni 21:55, 28 September 2008 (EEST) (type: truth; paradigms: science: attack)
What 'impossible' means may be ambiguous, because a clear probability limit cannot be given. The current attempt to define it is this: "If the probability that a range X contains the result of the variable does not deviate from zero in any practical sense, it is called 'impossible'." This means that if someone says that the value is possible, he has the burden of proof to show what difference, in some practical sense, it makes compared with the situation that the value is impossible and thus out of domain. This might be a problematic definition, as the practicality refers to the use purpose of the variable, and this may lead to circular inferences. On the other hand, it might be good to include the use purpose in the process of defining the domain. In any case, the result (being a probability distribution and thus giving weights to values in the domain) determines the impacts of the variable in a causal chain and in an assessment. The purpose of the domain is to 1) operationalise the work estimating the result and 2) make sure that a proper range of possible values is considered (which directly relates to the use purpose!). ←--4: . If inclusion is the default and exclusion needs argumentation, there will not be redundant discussion. See the example below. --Jouni 21:55, 28 September 2008 (EEST) (type: truth; paradigms: science: defence) It should be noted that impossibility is a stronger statement than possibility; if a range is found impossible, the conclusion is that ALL its values are impossible; if it is found possible, SOME of its values are found possible. Therefore, those who want to have a large domain, can start a discussion about including an unrealistically large domain. It must first be accepted, if it contains possible values, and then the impossible values must be excluded separately. This is in practice the same as taking the inclusion as the default in the first place. ←--5: . If no discussion on the domain has been performed, all values are possible. In the opposite case, all values are impossible until there has been at least one discussion about the domain. --Jouni 21:55, 28 September 2008 (EEST) (type: truth; paradigms: science: defence) |
Renaming needed?
| Fact discussion: . |
|---|
| Opening statement: The suggested new sub-attribute, now called domain, should be named more specifically
Closing statement: The original term 'domain' is changed to 'result domain'. (A closing statement, when resolved, should be updated to the main page.) |
| Argumentation:
←--1: . The word domain has several meanings (see e.g. Merriam-Webster's online dictionary), of which some could be confused with the meaning as used in the current explanation of the sub-attribute (which is along the lines of the meaning as the set of elements to which a mathematical or logical variable is limited ; specifically: the set on which a function is defined). For example the meaning along the lines of a sphere of knowledge, influence, or activity <the domain of art> could be considered as a relevant interpretation of the word domain in this context, but its operationalisation would probably be quite different within the information structure. A possible name for the sub-attribute as currently explained could be e.g. result domain. --Mikko Pohjola 18:50, 28 September 2008 (EEST) (type: truth; paradigms: science: defence) ⇤--2: . I agree with the reasoning in defend 1, and I also agree that 'result domain' could be used as a more specific term. However, so far I have not seen convincing evidence that such misunderstanding would occur in practice. The term should be as simple as possible. --Jouni 22:54, 28 September 2008 (EEST) (type: truth; paradigms: science: attack)
|