Thurstonian Item Response Theory and an Application to Attitude Items

  • Edward H. IpEmail author
Part of the Springer Proceedings in Mathematics & Statistics book series (PROMS, volume 89)


The assessment of attitudes has a long history dating back at least to the work of Thurstone. The Thurstonian approach had its “golden days,” but today it is seldom used, partly because judges are needed to assess the location of an item, but also because of the emergence of contemporary tools such as the IRT. The current work is motivated by a study that assesses medical students’ attitudes toward obese patients. During the item-development phase, the study team discovered that there were items on which the team members could not agree with regard to whether they represented positive or negative attitudes. Subsequently, a panel of n = 201 judges from the medical profession were recruited to rate the items, and the responses to the items were collected from a sample of n = 103 medical students. In the current work, a new methodology is proposed to extend the IRT for scoring student responses, and an affine transformation maps the judges’ scale onto the IRT scale. The model also takes into account measurement errors in the judges’ ratings. It is demonstrated that the linear logistic test model can be used to implement the proposed Thurstonian IRT approach.


Item response theory Likert scaling Linear logistic test model Attitudes toward obese persons Equal-appearing interval scaling 


  1. Allison DB, Basile VC, Yuker HE (1991) The measurement of attitudes toward and beliefs about obese persons. Int J Eat Disord 10:599–607Google Scholar
  2. Babbie ER (2008) The basics of social research, 4th edn. Thomson Learning, Inc., Belmont, CAGoogle Scholar
  3. Bergner M, Bobbitt RA, Carter WB, Gilson BS (1981) The sickness impact profile: development and final revision of a health status measure. Med Care 14:787–805CrossRefGoogle Scholar
  4. Brown A, Maydeu-Olivares A (2012) Fitting a Thurstonian IRT model to forced-choice data using Mplus. Behav Res Methods 44:1135–1147CrossRefGoogle Scholar
  5. Caffo B, An M, Rohde C (2007) Flexible random intercept model for binary outcomes using mixture of normals. Comput Stat Data Anal 51:5220–5235CrossRefzbMATHMathSciNetGoogle Scholar
  6. Coombs CH (1964) A theory of data. Wiley, New YorkGoogle Scholar
  7. Edwards AL, Kenney KC (1946) A comparison of the Thurstone and Likert techniques of attitude scale construction. J Appl Psychol 30:72–83CrossRefGoogle Scholar
  8. Fischer GA (1973) The linear logistic test model as an instrument in educational research. Acta Psychol 37:359–374CrossRefGoogle Scholar
  9. Ip EH (2010) Empirically indistinguishable multidimensional IRT and locally dependent unidimensional item response models. Br J Math Stat Psychol 63:395–415CrossRefMathSciNetGoogle Scholar
  10. Ip EH, Smits D, De Boeck P (2009) Locally dependent linear logistic test model with person covariates. Appl Psychol Meas 33:555–569CrossRefGoogle Scholar
  11. Ip EH, Marshall S, Crandall SJ, Vitolins M, Davis S, Miller D, Kronner D, Vaden K, Spangler J (2013) Measuring medical student attitudes and beliefs regarding obese patients. Acad Med 88:282–289CrossRefGoogle Scholar
  12. Lewis RJ, Cash TF, Jacobi L, Bubb-Lewis C (1997) Prejudice toward fat people: the development and validation of the antifat attitudes test. Obes Res 5:297–307CrossRefGoogle Scholar
  13. Likert R (1932) A technique for the measurement of attitudes. Arch Psychol 22(140):1–55Google Scholar
  14. Lord FM (1980) Applications of item response theory to practical testing problems. Erlbaum, Hillsdale, NJGoogle Scholar
  15. Mair P, Hatzinger R (2007) Extended Rasch modeling: the eRm package for application of IRT models in R. J Stat Softw 20(9). Last assessed December 11th, 2013.
  16. Massof RW (2002) The measurement of vision disability. Optom Vis Sci 79:516–552CrossRefGoogle Scholar
  17. Ostram TM (1989) Interdependence of attitude theory and measurement. In: Pratkanis AR, Breckler SJ, Greenwald AG (eds) Attitude structure and function. Lawrence Erlbaum Associates, Hillsdale, NJ, pp 11–36Google Scholar
  18. Roberts JS, Laughlin JE (1996) A unidimensional item response model for unfolding responses from a graded disagree–agree response scale. Appl Psychol Meas 20:231–255CrossRefGoogle Scholar
  19. Thurstone LL (1925) A method of scaling psychological and educational tests. J Educ Psychol 16:433–451CrossRefGoogle Scholar
  20. Thurstone LL (1928) Attitudes can be measured. Am Coll Sociol 33:529–554CrossRefGoogle Scholar
  21. Thurstone LL, Chave EJ (1929) The measurement of social attitudes. University of Chicago Press, ChicagoGoogle Scholar
  22. Williams VSL, Pommerich M, Thissen D (1998) A comparison of developmental scales based on Thurstone methods and item response theory. J Educ Meas 35:93–107CrossRefGoogle Scholar
  23. Zeger SL, Liang KY, Albert P (1988) Models for longitudinal data: a generalized estimating equation approach. Biometrics 44:1049–1060CrossRefzbMATHMathSciNetGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  1. 1.Department of Biostatistical SciencesWake Forest School of MedicineWinston-SalemUSA

Personalised recommendations