Skip to main content

Priors in Bayesian Estimation Under the Two-Parameter Logistic Model

  • Conference paper
  • First Online:
Quantitative Psychology (IMPS 2020)

Abstract

A review of various priors used in Bayesian estimation under the two-parameter logistic model is presented together with clear mathematical definitions of the prior distributions. Examples that compared Bayesian estimation methods are presented using empirical data. The effects of the priors and their specifications on both item and ability parameter estimates are demonstrated. The computer program OpenBUGS that implements the rejection sampling method is the main program employed in the study. Issues in Bayesian estimation and use of priors in item response theory are discussed.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 169.00
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 219.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info
Hardcover Book
USD 219.99
Price excludes VAT (USA)
  • Durable hardcover edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others

References

  • Albert, J. H. (1992). Bayesian estimation of normal ogive item response curves using Gibbs sampling. Journal of Educational Statistics, 17, 251–269.

    Article  Google Scholar 

  • Birnbaum, A. (1969). Statistical theory for logistic mental test models with a prior distribution of ability. Journal of Mathematical Psychology, 6, 258–276.

    Article  MATH  Google Scholar 

  • Bock, R. D., & Aitkin, M. (1981). Marginal maximum likelihood estimation of item parameters: Applications of an EM algorithm. Psychometrika, 46, 443–459.

    Article  MathSciNet  Google Scholar 

  • Bock, R. D., & Lieberman, M. (1970). Fitting a response model for n dichotomously scored items. Psychometrika, 35, 179–197.

    Article  Google Scholar 

  • du Toit, M. (Ed.). (2003). IRT from SSI: BILOG-MG, MULTILOG, PARSCALE, TESTFACT. Chicago, IL: Scientific Software International.

    Google Scholar 

  • Fox, J.-P. (2010). Bayesian item response modeling: Theory and applications. New York, NY: Springer.

    Book  MATH  Google Scholar 

  • Gonzalez, J. (2010). Bayesian methods in psychological research: The case of IRT. International Journal of Psychological Research, 3(1), 164–176.

    Google Scholar 

  • Isaacs, G. I., Christ, D. E., Novick, M. R., & Jackson, P. H. (1974). Tables for Bayesian statisticians. The Iowa Testing Program, The University of Iowa, Iowa City, IA.

    MATH  Google Scholar 

  • Johnson, M. S., & Sinharay, S. (2016). Bayesian etimation. In W. J. van der Linden (Ed.), Handbook of item response theory, Volume 2: Statistical tools (pp. 237–257). Boca Raton, FL: CRC Press.

    Google Scholar 

  • Kim, J.-S., & Bolt, D. M. (2007). Estimating item response theory models using Markov chain Monte Carlo methods. Educational Measurement: Issues and Practice, 26(4), 38–51.

    Article  Google Scholar 

  • Kim, S.-H. (2001). An evaluation of a Markov chain Monte Carlo method for the Rasch model. Applied Psychological Measurement, 25, 163–176.

    Article  MathSciNet  Google Scholar 

  • Kim, S.-H., Cohen, A. S., Baker, F. B., Subkoviak, M. J., & Leonard, T. (1994). An investigation of hierarchical Bayes procedures in item response theory. Psychometrika, 59, 405–421.

    Article  MATH  Google Scholar 

  • Leonard, T., & Novick, M. R. (1985). Bayesian inference and diagnostics for the three parameter logistic model (ONR Technical Report Np. 85-5). Iowa City, IA: The University of Iowa, Cada Research Group. (ERIC Document Reproduction Service No. ED261068).

    Google Scholar 

  • Levy, R., & Mislevy, R. J. (2016). Bayesian psychometric modeling. Boca Raton, FL: CRC Press.

    MATH  Google Scholar 

  • Lindley, D. V., & Smith, A. F. (1972). Bayesian estimates for the linear model. Journal of the Royal Statistical Society, Series B, 34, 1–41.

    MathSciNet  MATH  Google Scholar 

  • Lunn, D., Jackson, C., Best, N., Thomas, A., & Spiegelhalter, D. (2013). The BUGS book: A practical introduction the Bayesian analysis. Boca Raton, FL: CRC Press.

    MATH  Google Scholar 

  • Marcoulides, K. M. (2018). Careful with those priors: A note on Bayesian estimation in two-parameter logistic item response theory models. Measurement, 16, 92–99.

    Google Scholar 

  • Mislevy, R. J. (1986). Bayes modal estimation in item response models. Psychometrika, 51, 177–195.

    Article  MathSciNet  MATH  Google Scholar 

  • Novick, M. R., & Jackson, P. H. (1974). Statistical methods for educational and psychological research. New York, NY: McGraw-Hill.

    Google Scholar 

  • Novick, M. R., Lewis, C., & Jackson, P. H. (1973). The estimation of proportions in n groups. Psychometrika, 38, 19–46.

    Article  MathSciNet  MATH  Google Scholar 

  • Spiegelhalter, D., Thomas, A., Best, N., & Lunn, D. (2014). OpenBUGS user manual. Cambridge, UK: MRC Biostatistics Unit, Institute of Public Health.

    Google Scholar 

  • Stone, C. A., & Zhu, X. (2015). Bayesian analysis of item response theory models using SAS. Cary, NC: SAS Institute.

    Google Scholar 

  • Swaminathan, H., & Gifford, J. A. (1982). Bayesian estimation in the Rasch model. Journal of Educational Statistics, 7, 175–191.

    Article  Google Scholar 

  • Swaminathan, H., & Gifford, J. A. (1985). Bayesian estimation in the two-parameter logistic model. Psychometrika, 50, 349–364.

    Article  MATH  Google Scholar 

  • Swaminathan, H., & Gifford, J. A. (1986). Bayesian estimation in the three-parameter logistic model. Psychometrika, 51, 589–601.

    Article  MathSciNet  MATH  Google Scholar 

  • Thissen, D., Chen, W.-H., & Bock, R. D. (2002). MULTILOG [Computer software]. Lincolnwood, IL: Scientific Software International.

    Google Scholar 

  • Tsutakawa, R. K., & Lin, H. Y. (1986). Bayesian estimation of item response curves. Psychometrika, 51, 251–267.

    Article  MathSciNet  MATH  Google Scholar 

  • Wright, B. D., & Stone, M. H. (1979). Best test design. Chicago, IL: MESA Press.

    Google Scholar 

  • Zimowski, M. F., Muraki, E., Mislevy, R. J., & Bock, R. D. (2002). BILOG-MG [Computer software]. Lincolnwood, IL: Scientific Software International.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Seock-Ho Kim .

Editor information

Editors and Affiliations

Appendices

Appendix A: OpenBUGS Code

model { # 2PL model   for (i in 1:I) {     for (j in 1:J) {       logit(p[i, j]) <- alpha[j] ∗ (theta[i] - beta[j])       x[i, j] ~ dbern(p[i, j])     } # ability prior     theta[i] ~ dnorm(0, 1)   } # item Priors   for (j in 1:J) {     a[j] ~ dchisqr(10)     alpha[j] <- sqrt(a[j] ∗ 0.1)     beta[j] ~ dunif(-5, 5) #   beta[j] ~ dnorm(mub, taub) # GS2 #   alpha[j] ~ dlnorm(0, 2) # GS3 #   beta[j] ~ dnorm(0, 0.5) # GS3   } # hyperpriors # mub ~ dflat() # GS2 # taub ~ dgamma(2.5, 5) # GS2 } # kct data list(I = 35, J = 18, x = structure(.Data = c( 1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,0,0,1,0,0,0,0,0,0, 1,1,1,1,0,0,1,0,1,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,1,1,0,1,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0, 1,1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,0,1,0,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,0,0,1,1,1,1,1,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,0,1,0,0,0,0,0,0,0, 1,1,1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,1,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,1,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,0,1,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,1,1,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,0,1,0,0,1,1,0, 1,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,0,1,0,0,0,0,0,0,0, 1,1,1,1,1,1,0,0,1,1,1,0,0,1,0,0,0,0, 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0, 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,0,1,0,1,0,0,0,0, 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ), .Dim = c(35, 18)) ) # initial values (e.g., GS1) list( a = c( 10,10,10,10,10,10, 10,10,10,10,10,10, 10,10,10,10,10,10 ), beta = c( 0,0,0,0,0,0, 0,0,0,0,0,0, 0,0,0,0,0,0 ),

#  mub = 0, taub = 1, theta = c( -0.4519851,  0.2231436,  0.2231436, -0.6931472,  0.2231436,  0.2231436, 1.2527630,  0.2231436,  0.2231436,  0.4519851, -0.2231436, -0.2231436, 0.2231436,  0.4519851,  0.9555114,  0.2231436,  0.0000000,  0.4519851, 0.0000000,  0.4519851,  0.6931472,  0.6931472,  0.6931472,  1.2527630, -0.9555114,  0.2231436, -0.4519851,  0.2231436,  0.2231436,  0.0000000, 0.2231436,  0.4519851, -0.6931472,  0.6931472, -1.6094379 ) )

Appendix B: Summary of Priors and Specifications

Papers in Tables 4, 5 and 6 are not exhaustive. Estimation techniques in the tables include JBME, MCMC, and MBE. The acronym BME designates Bayes modal estimation, BE designates Bayes estimation (i.e., posterior mean), EAP designates expected a posteriori (i.e., posterior mean via quadratures), and MAP designates maximum a posteriori (i.e., posterior mode with known item parameters). The types of priors can be classified into two; one without any hierarchical structure and the other with some hierarchical structure for which parameters are modeled with hyperpriors and hyperparameters (i.e., Hierarchical). Priors can also be differentiated as ones with exchangeability for which the same prior will be applied to all items in a test or a subtest (i.e., Exchangeable), others with capability of assigning an individual prior on each parameter (i.e., Individual), and also others obtained with information from the current data (i.e., Empirical). It should be noted that in the tables, the names of the distributions might sound the same but could be mathematically, trivially different. Each paper should be consulted and carefully read before employing the priors in one’s research. Also note that several keywords from the computer programs (e.g., SPR, TPR, FLO, AJ, BJ, PA, etc.) are used without any explications.

There are more than six additional, relevant papers that could be included in Tables 4, 5 and 6. The relevant papers are as follows (but without full references): Spiegelhalter et al.’s (1996) “BUGS 0.5 Examples Volume 1”; Johnson and Albert’s (1999) “Ordinal Data Modeling”; Curtis’s (2010) “Journal of Statistical Software, 36”; Nathesan et al.’s (2016) “Frontiers in Psychology, 7”; Luo and Ziao’s (2017) “Educational and Psychological Measurement, Febuary 1”; and Parchev et al.’s (2017) “CRAN Package irtoys”. The six papers mentioned in Tables 4, 5 and 6 are representative ones.

Rights and permissions

Reprints and permissions

Copyright information

© 2021 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Kim, SH. et al. (2021). Priors in Bayesian Estimation Under the Two-Parameter Logistic Model. In: Wiberg, M., Molenaar, D., González, J., Böckenholt, U., Kim, JS. (eds) Quantitative Psychology. IMPS 2020. Springer Proceedings in Mathematics & Statistics, vol 353. Springer, Cham. https://doi.org/10.1007/978-3-030-74772-5_28

Download citation

Publish with us

Policies and ethics