Quantifying definitional relations in psychological measurement
(2020) In New Ideas in Psychology 56.- Abstract
Quantitative inferences about psychological attributes, such as extraversion, depression and empathy, involve measurement instruments as well as mathematical models that specify how indicators should be aggregated. The type of model that is appropriate for doing so is conditional on the type of relation that exists between an attribute, its facets and its indicators. The common assumption is that such relations are causal relations. Here, instead, we address definitional attribute-facet relations. Our aim is to find an appropriate mathematical model for when an attribute is defined in terms of facets, instead of causing or being caused by facets. In doing so, we describe the semantics of definitions in logical form. From this form we... (More)
Quantitative inferences about psychological attributes, such as extraversion, depression and empathy, involve measurement instruments as well as mathematical models that specify how indicators should be aggregated. The type of model that is appropriate for doing so is conditional on the type of relation that exists between an attribute, its facets and its indicators. The common assumption is that such relations are causal relations. Here, instead, we address definitional attribute-facet relations. Our aim is to find an appropriate mathematical model for when an attribute is defined in terms of facets, instead of causing or being caused by facets. In doing so, we describe the semantics of definitions in logical form. From this form we then derive continuous functions for attribute-facet relations using fuzzy logic. A model with main effects and interactions between facets seems to be more powerful for representing definitional relations than traditional formative and reflective models. This has important implications for measurement in basic and applied research.
(Less)
- author
- Rohner, Jean Christophe LU and Levinsson, Henrik LU
- organization
- publishing date
- 2020-01
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Causal relation, Construct validity, Definitional relation, Fuzzy logic, Measurement model, Psychological assessment
- in
- New Ideas in Psychology
- volume
- 56
- article number
- 100749
- publisher
- Elsevier
- external identifiers
-
- scopus:85080837736
- ISSN
- 0732-118X
- DOI
- 10.1016/j.newideapsych.2019.100749
- language
- English
- LU publication?
- yes
- id
- 255f7967-9162-4f49-8d8f-2ad57e353c9b
- date added to LUP
- 2020-03-20 15:38:35
- date last changed
- 2022-04-18 21:24:16
@article{255f7967-9162-4f49-8d8f-2ad57e353c9b, abstract = {{<p>Quantitative inferences about psychological attributes, such as extraversion, depression and empathy, involve measurement instruments as well as mathematical models that specify how indicators should be aggregated. The type of model that is appropriate for doing so is conditional on the type of relation that exists between an attribute, its facets and its indicators. The common assumption is that such relations are causal relations. Here, instead, we address definitional attribute-facet relations. Our aim is to find an appropriate mathematical model for when an attribute is defined in terms of facets, instead of causing or being caused by facets. In doing so, we describe the semantics of definitions in logical form. From this form we then derive continuous functions for attribute-facet relations using fuzzy logic. A model with main effects and interactions between facets seems to be more powerful for representing definitional relations than traditional formative and reflective models. This has important implications for measurement in basic and applied research.</p>}}, author = {{Rohner, Jean Christophe and Levinsson, Henrik}}, issn = {{0732-118X}}, keywords = {{Causal relation; Construct validity; Definitional relation; Fuzzy logic; Measurement model; Psychological assessment}}, language = {{eng}}, publisher = {{Elsevier}}, series = {{New Ideas in Psychology}}, title = {{Quantifying definitional relations in psychological measurement}}, url = {{http://dx.doi.org/10.1016/j.newideapsych.2019.100749}}, doi = {{10.1016/j.newideapsych.2019.100749}}, volume = {{56}}, year = {{2020}}, }