Satisfaction Ratings of QOLPAV: Psychometric Properties Based on the Graded Response Model

Ssu Kuang Chen, Fang Ming Hwang*, San-Ju Lin

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

4 Scopus citations


A scale measuring quality of life (QOL) is important in adolescent research. Using the graded response model (GRM), this study evaluates the psychometric properties of the satisfaction ratings of the Quality of Life Profile Adolescent Version (QOLPAV). Data for 1,392 adolescents were used to check IRT assumptions such as unidimensionality and local item dependence (LID). The goodness of fit of the GRM to the data and the item characteristic curves were evaluated. The reliability and validity analyses included item/test information, Cronbach's α, and convergent and discriminant validity. Differential item functioning (DIF) procedures were also performed to detect item bias. The results provide evidence that the items sufficiently measured one single dimension. Few pairs of questions were flagged as LID due to content or wording similarity. Five items did not fit the GRM, and 4 were low in item discrimination. The findings also suggest that the assessment had appropriate reliability and validity. The DIF impact on the assessment score was considered minor. Because QOLPAV includes a respondent's perceived importance of various life aspects, a short form that only considers important life aspects in the overall QOL estimation for each respondent becomes feasible within the framework of IRT. Future studies focusing on the development of a QOL overall index using the items from QOLPAV are recommended.

Original languageEnglish
Pages (from-to)367-383
Number of pages17
JournalSocial Indicators Research
Issue number1
StatePublished - 1 Jan 2013


  • Importance weights
  • Pyschometric properties
  • Quality of life
  • Satisfaction ratings
  • The graded response model

Fingerprint Dive into the research topics of 'Satisfaction Ratings of QOLPAV: Psychometric Properties Based on the Graded Response Model'. Together they form a unique fingerprint.

Cite this