Normal view MARC view ISBD view

Logistic and linear models as a framework for the analysis of the modified version of Social Participation Questionnaire / Una Mikac, Andrea Hildebrandt, Vesna Buško, Laura Kaltwasser.

By: Mikac, Una.
Contributor(s): Hildebrandt, Andrea [aut] | Buško, Vesna [aut] | Kaltwasser, Laura [aut].
Material type: ArticleArticleDescription: 52-52 str.Other title: Logistic and linear models as a framework for the analysis of the modified version of Social Participation Questionnaire [Naslov na engleskom:].Subject(s): 5.06 | Social Participation Questionnaire, EFA, CFA, Rasch, dimensionality | Social Participation Questionnaire, EFA, CFA, Rasch, dimensionality In: 23. Dani Ramira i Zorana Bujasa str. 52-52Abstract: Test developers use different psychometric models depending on the theoretical background of the scale to be developed. The essential difference for the model choice relies on the theoretical assumption about the nature of a scale behind recorded answers. If the recorded answers are polytomous, but no continuous scale can be assumed behind recorded responses, in an item response theoretical framework logistic relation between item and trait is to be modeled. However, if polytomous answers can be assumed to thresholds along a continuous scale, confirmatory factor models can be used to model the relation between item and trait. We will compare information attained from the use of different psychometric modeling approaches in test development with a focus on the consolidation of the different approaches regarding the structure of the data. The analyses will be performed on responses of 219 participants on the 30-item Social Participation Questionnaire, adapted and translated to German. Social participation contributes to quality of life (Baum et al, 2000) and could be used as an indicator of social activity. The instrument was first developed in context of Rasch theory with the presumed unidimensional structure, although groups of items based on content were recognized based on expert knowledge (Densley, Davidson, & Gunn, 2013). The authors of the questionnaire stressed the need to explore the facets of social participation. Our data indicate two higher-order factors when using factor analysis, while principal component analysis of Rasch residuals does not show strong deviations from unidimensionality. Therefore, we performed the item analysis on the whole questionnaire and on the two subscales using both logistic and linear models. The influence of different outcomes of the analyses on the decisions on the final version of the questionnaire will be discussed in the light of different theoretical backgrounds, with practical implications for future instrument development.
Tags from this library: No tags from this library for this title. Log in to add tags.
No physical items for this record

Test developers use different psychometric models depending on the theoretical background of the scale to be developed. The essential difference for the model choice relies on the theoretical assumption about the nature of a scale behind recorded answers. If the recorded answers are polytomous, but no continuous scale can be assumed behind recorded responses, in an item response theoretical framework logistic relation between item and trait is to be modeled. However, if polytomous answers can be assumed to thresholds along a continuous scale, confirmatory factor models can be used to model the relation between item and trait. We will compare information attained from the use of different psychometric modeling approaches in test development with a focus on the consolidation of the different approaches regarding the structure of the data. The analyses will be performed on responses of 219 participants on the 30-item Social Participation Questionnaire, adapted and translated to German. Social participation contributes to quality of life (Baum et al, 2000) and could be used as an indicator of social activity. The instrument was first developed in context of Rasch theory with the presumed unidimensional structure, although groups of items based on content were recognized based on expert knowledge (Densley, Davidson, & Gunn, 2013). The authors of the questionnaire stressed the need to explore the facets of social participation. Our data indicate two higher-order factors when using factor analysis, while principal component analysis of Rasch residuals does not show strong deviations from unidimensionality. Therefore, we performed the item analysis on the whole questionnaire and on the two subscales using both logistic and linear models. The influence of different outcomes of the analyses on the decisions on the final version of the questionnaire will be discussed in the light of different theoretical backgrounds, with practical implications for future instrument development.

Projekt MZOS projekt

ENG

There are no comments for this item.

Log in to your account to post a comment.

Powered by Koha

//