Donna Harrington
- Published in print:
- 2008
- Published Online:
- January 2009
- ISBN:
- 9780195339888
- eISBN:
- 9780199863662
- Item type:
- book
- Publisher:
- Oxford University Press
- DOI:
- 10.1093/acprof:oso/9780195339888.001.0001
- Subject:
- Social Work, Research and Evaluation
Measures that are reliable, valid, and can be used across diverse populations are vital to social work research, but the development of new measures is an expensive and time-consuming process. An ...
More
Measures that are reliable, valid, and can be used across diverse populations are vital to social work research, but the development of new measures is an expensive and time-consuming process. An array of existing measures can provide a cost-effective alternative, but in order to take this expedient step with confidence, researchers must ensure that the existing measure is appropriate for the new study. Confirmatory factor analysis (CFA) is one of the ways to do so. CFA has four primary functions—psychometric evaluation of measures, construct validation, testing method effects, and testing measurement invariance. This book provides an overview of the method, step-by-step guides to creating a CFA model and assessing its fit, and explanations of the requirements for using CFA, as well the book underscores the issues that are necessary to consider when using multiple groups or equivalent and multilevel models. Real-world examples, screenshots from the Amos software program that can be used to conduct CFA, and reading suggestions for each chapter form part of the book.Less
Measures that are reliable, valid, and can be used across diverse populations are vital to social work research, but the development of new measures is an expensive and time-consuming process. An array of existing measures can provide a cost-effective alternative, but in order to take this expedient step with confidence, researchers must ensure that the existing measure is appropriate for the new study. Confirmatory factor analysis (CFA) is one of the ways to do so. CFA has four primary functions—psychometric evaluation of measures, construct validation, testing method effects, and testing measurement invariance. This book provides an overview of the method, step-by-step guides to creating a CFA model and assessing its fit, and explanations of the requirements for using CFA, as well the book underscores the issues that are necessary to consider when using multiple groups or equivalent and multilevel models. Real-world examples, screenshots from the Amos software program that can be used to conduct CFA, and reading suggestions for each chapter form part of the book.
Ralph B. Taylor
- Published in print:
- 2015
- Published Online:
- March 2016
- ISBN:
- 9780814725498
- eISBN:
- 9780814708033
- Item type:
- chapter
- Publisher:
- NYU Press
- DOI:
- 10.18574/nyu/9780814725498.003.0008
- Subject:
- Sociology, Law, Crime and Deviance
This chapter highlights two conceptual roadblocks to systematic comparisons of competing ecological communities and crime models. First, key indicators in several communities and crime models are ...
More
This chapter highlights two conceptual roadblocks to systematic comparisons of competing ecological communities and crime models. First, key indicators in several communities and crime models are plagued by semantic ambiguity. In essence, different researchers connect the same indicators to different constructs. Second, researchers rarely engage in systematic, multimethod ecological construct validation to clarify which indicators clearly belong to which constructs. Multimethod patterns of convergent and discriminant validation are rarely examined. Of course, such examinations are particularly challenging when examining spatially aggregated indicators. The establishment of discriminant validity, a key part of the construct validation process, is needed but challenging to accomplish with data aggregated by geographic proximity. The solution is to first adopt a unified perspective on the construct validation process, which opens up the range of empirical relationships relevant to establishing construct validity. The second is to carry out the unified construct validation activities using a Boudon/Coleman boat metamodel aligned with methodological individualism.Less
This chapter highlights two conceptual roadblocks to systematic comparisons of competing ecological communities and crime models. First, key indicators in several communities and crime models are plagued by semantic ambiguity. In essence, different researchers connect the same indicators to different constructs. Second, researchers rarely engage in systematic, multimethod ecological construct validation to clarify which indicators clearly belong to which constructs. Multimethod patterns of convergent and discriminant validation are rarely examined. Of course, such examinations are particularly challenging when examining spatially aggregated indicators. The establishment of discriminant validity, a key part of the construct validation process, is needed but challenging to accomplish with data aggregated by geographic proximity. The solution is to first adopt a unified perspective on the construct validation process, which opens up the range of empirical relationships relevant to establishing construct validity. The second is to carry out the unified construct validation activities using a Boudon/Coleman boat metamodel aligned with methodological individualism.
Catherine S. Taylor
- Published in print:
- 2013
- Published Online:
- March 2015
- ISBN:
- 9780199791040
- eISBN:
- 9780190256074
- Item type:
- chapter
- Publisher:
- Oxford University Press
- DOI:
- 10.1093/acprof:osobl/9780199791040.003.0005
- Subject:
- Psychology, Social Psychology
This chapter deals with construct-related evidence for the validity of test scores. Evaluating construct-related evidence for validity involves evaluating the logical arguments and empirical evidence ...
More
This chapter deals with construct-related evidence for the validity of test scores. Evaluating construct-related evidence for validity involves evaluating the logical arguments and empirical evidence supporting the central claim (implied or explicit) in assessment—that the scores from an assessment can be interpreted and used in a particular way. Score interpretations may include predictions about examinees' most likely performance on a criterion (e.g. piloting an airplane or success in college) or to make inferences about examinees' measures on internal constructs (e.g. depression, reading comprehension, attitudes toward science). Evidence is needed to determine whether these inferences can be made.Less
This chapter deals with construct-related evidence for the validity of test scores. Evaluating construct-related evidence for validity involves evaluating the logical arguments and empirical evidence supporting the central claim (implied or explicit) in assessment—that the scores from an assessment can be interpreted and used in a particular way. Score interpretations may include predictions about examinees' most likely performance on a criterion (e.g. piloting an airplane or success in college) or to make inferences about examinees' measures on internal constructs (e.g. depression, reading comprehension, attitudes toward science). Evidence is needed to determine whether these inferences can be made.
Catherine S. Taylor
- Published in print:
- 2013
- Published Online:
- March 2015
- ISBN:
- 9780199791040
- eISBN:
- 9780190256074
- Item type:
- chapter
- Publisher:
- Oxford University Press
- DOI:
- 10.1093/acprof:osobl/9780199791040.003.0006
- Subject:
- Psychology, Social Psychology
This chapter focuses on the types of research needed to support (or refute) interpretations and uses of test scores. It begins by discussing studies to investigate whether there is support for score ...
More
This chapter focuses on the types of research needed to support (or refute) interpretations and uses of test scores. It begins by discussing studies to investigate whether there is support for score interpretations. It then presents examples of research for investigating score uses. Finally, it considers the social consequences of test score interpretations and uses.Less
This chapter focuses on the types of research needed to support (or refute) interpretations and uses of test scores. It begins by discussing studies to investigate whether there is support for score interpretations. It then presents examples of research for investigating score uses. Finally, it considers the social consequences of test score interpretations and uses.
Anna Alexandrova
- Published in print:
- 2017
- Published Online:
- August 2017
- ISBN:
- 9780199300518
- eISBN:
- 9780190677213
- Item type:
- chapter
- Publisher:
- Oxford University Press
- DOI:
- 10.1093/oso/9780199300518.003.0005
- Subject:
- Philosophy, Moral Philosophy
Among the scientists and policymakers, measurability of well-being is now almost uncontroversial—only the individual measures are. Philosopher Daniel Hausman, however, argues that well-being in the ...
More
Among the scientists and policymakers, measurability of well-being is now almost uncontroversial—only the individual measures are. Philosopher Daniel Hausman, however, argues that well-being in the proper sense is too person-relative and hence heterogeneous. It is not a quantity that can be measured on the population level. This chapter agrees that there is a sense of well-being—the all-things-considered individual well-being—on which it is likely not measurable but disagrees that therefore well-being is not measurable in any sense. Crucial to measurement is the existence of generalisations between core components of well-being and observable indicators. Such generalisations are available if well-being is predicated of kinds of people, rather than of individual lives. Assuming this focus on well-being of kinds, validity of most existing measures of well-being is secured by the process of construct validation, whose logic relies on a plausible ideal of balancing all evidence.Less
Among the scientists and policymakers, measurability of well-being is now almost uncontroversial—only the individual measures are. Philosopher Daniel Hausman, however, argues that well-being in the proper sense is too person-relative and hence heterogeneous. It is not a quantity that can be measured on the population level. This chapter agrees that there is a sense of well-being—the all-things-considered individual well-being—on which it is likely not measurable but disagrees that therefore well-being is not measurable in any sense. Crucial to measurement is the existence of generalisations between core components of well-being and observable indicators. Such generalisations are available if well-being is predicated of kinds of people, rather than of individual lives. Assuming this focus on well-being of kinds, validity of most existing measures of well-being is secured by the process of construct validation, whose logic relies on a plausible ideal of balancing all evidence.
Ralph B. Taylor
- Published in print:
- 2015
- Published Online:
- March 2016
- ISBN:
- 9780814725498
- eISBN:
- 9780814708033
- Item type:
- chapter
- Publisher:
- NYU Press
- DOI:
- 10.18574/nyu/9780814725498.003.0001
- Subject:
- Sociology, Law, Crime and Deviance
This chapter begins by discussing criminological scholars' concerns about the lack of integration across levels of theorizing, i.e. the failure of criminological theory to understand the connections ...
More
This chapter begins by discussing criminological scholars' concerns about the lack of integration across levels of theorizing, i.e. the failure of criminological theory to understand the connections between macrosocial, microsocial, and individual factors. It then sets out the book's main purpose, which is to describe four issues which, because they have been overlooked, have blocked theoretical integration in community criminology. It links these issues to key matters outside of criminology and provides an integrated framework for considering the four concerns. These four issues are: spatial scaling, temporal scaling, construct validation, and selection effects. The remainder of chapter provides definitions of terms used in the study, an introduction to the Boudon-Coleman “boat” metamodel, and an overview of the subsequent chapters.Less
This chapter begins by discussing criminological scholars' concerns about the lack of integration across levels of theorizing, i.e. the failure of criminological theory to understand the connections between macrosocial, microsocial, and individual factors. It then sets out the book's main purpose, which is to describe four issues which, because they have been overlooked, have blocked theoretical integration in community criminology. It links these issues to key matters outside of criminology and provides an integrated framework for considering the four concerns. These four issues are: spatial scaling, temporal scaling, construct validation, and selection effects. The remainder of chapter provides definitions of terms used in the study, an introduction to the Boudon-Coleman “boat” metamodel, and an overview of the subsequent chapters.
