Original article
Test for item bias in a quality of life questionnaire

https://doi.org/10.1016/0895-4356(94)00195-VGet rights and content

Abstract

Item bias (differential item functioning) analysis examines whether the construction of an index from two or more variables results in bias in relation to sex, age, or other criteria. Item bias may lead to erroneous conclusions because of distortion or dilution of the effects measured. In comparing groups, item bias analysis, tests whether the information about possible differences between groups, obtained by the variables constituting an index, are correctly passed on by the index score. We examined a quality of life questionnaire answered by 1189 breast cancer patients. We found age-bias or bias in the comparison of groups receiving different treatments in three out of nine indexes. Recommendations for the interpretation of these indexes are made. Item bias analysis is a useful method examining an issue not covered by traditional psychometric tests.

References (25)

  • O Olsen et al.

    Age and the operationalization of social support

    Soc Sci Med

    (1991)
  • A Bowling
  • I McDowell et al.
  • SJ Osterlind
  • Bjorner JB, Groenvold M, Kreiner S. Item bias in medical indexes—concepts, techniques, and implications. J Clin...
  • S Kreiner

    Validation of index scales for analysis of survey data: the Symptom Index

  • JC Nunnally et al.
  • K Avlund et al.

    The measurement of Instrumental ADL: Content validity and construct validity

    Aging Clin Exp Res

    (1993)
  • DBCG-89. Program for behandling og kontrol af patienter med primær, operabel cancer mammae (DBCG-89. Program for treatment and follow-up of patients with primary, operable breast cancer)

    (1989)
  • Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. 133 randomised trials involving 31,000 recurrences and 24,000 deaths among 75,000 women

    Lancet

    (1992)

    Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. 133 randomised trials involving 31,000 recurrences and 24,000 deaths among 75,000 women

    Lancet

    (1992)
  • Cited by (57)

    • The calibrated, unidimensional anxiety item bank for cardiovascular patients provided the basis for anxiety assessment in cardiovascular rehabilitation patients

      2013, Journal of Clinical Epidemiology
      Citation Excerpt :

      It occurs when subgroups (e.g., women and men) answer differently to an item, although they have an identical underlying level of anxiety [42,48]. The analyses of DIF were calculated for four sociodemographic variables (gender, age, educational level, and employment status) and three disease-specific variables (intensity of pain, subjective limitations because of CVD, and specific cardiovascular diagnoses), which have been shown to be significant confounders of the association between anxiety and somatic diseases [49–52]. The existence of DIF was calculated by an analysis of variance of the person–item deviation residuals with person factors and class intervals as factors [42,47].

    • Development of a PROMIS item bank to measure pain interference

      2010, Pain
      Citation Excerpt :

      Analysis of differential item functioning (DIF) examines the relationships among item responses, levels of the domain being measured, and subgroup membership. For any given level of domain, the probabilities of endorsing specified item responses should be independent of subgroup membership [23]. In the context of pain interference measurement, persons of different ages, education levels, race/ethnicities, and genders who have equal levels of pain interference should be equally likely to endorse a particular category of a specified PROMIS-PI item.

    • Item response theory was used to shorten EORTC QLQ-C30 scales for use in palliative care

      2006, Journal of Clinical Epidemiology
      Citation Excerpt :

      We used the contingency table method described previously [18] to test for DIF between subgroups defined by age, gender, cancer site, language, and stage (i.e., palliative vs. nonpalliative). For a more detailed description of the DIF analysis method, see Petersen et al. [18]; see also Groenvold et al. [19] and Kreiner et al. [20]. The effect of possible DIF findings with regard to age, gender, and cancer site were evaluated by dividing the palliative care patients into subgroups according to DIF and then comparing the properties of the shortened scales in these subgroups.

    View all citing articles on Scopus
    View full text