MTP2 and Partial Correlations in Monotone Higher-Order Factor Models

  • Jules L. EllisEmail author
Part of the Springer Proceedings in Mathematics & Statistics book series (PROMS, volume 89)


For binary variables, multivariate positivity of order 2 (MTP2) implies nonnegative partial correlations (NPC). This is so because for any triple of variables, MTP2 is equivalent with conditional association.

Under weak distribution assumptions of the noise variables, monotone higher-order one-factor models imply MTP2 of the manifest variables. This remains true after discretization of the manifest variables. Therefore, MTP2 and NPC cannot be used to discriminate unidimensional monotone latent variable models from multidimensional monotone higher-order one-factor models.


Conditional association Multivariate positivity of order 2 Nonlinear factor analysis Partial correlation Second-order factor Supermodularity 


  1. Bartolucci F, Forcina A (2005) Likelihood inference on the underlying structure of IRT models. Psychometrika 70:31–43CrossRefzbMATHMathSciNetGoogle Scholar
  2. Chen FF, Sousa KH, West SG (2005) Testing measurement invariance of second-order factor models. Struct Equ Model 12:471–491CrossRefMathSciNetGoogle Scholar
  3. Denuit M, Dhaene J, Goovaerts M, Kaas R (2005) Actuarial theory for dependent risks. Wiley, ChichesterCrossRefGoogle Scholar
  4. De Gooijer JG, Yuan A (2011) Some exact tests for manifest properties of latent trait models. Comput Stat Data Anal 55:34–44CrossRefzbMATHGoogle Scholar
  5. Efron B (1965) Increasing properties of Pólya frequency functions. Ann Math Stat 36:272–279CrossRefzbMATHMathSciNetGoogle Scholar
  6. Ellis JL (2014) An inequality for correlations in unidimensional monotone latent variable models for binary variables. Psychometrika 79:303–316. doi: 10.1007/S11336-013-9341-5 CrossRefzbMATHMathSciNetGoogle Scholar
  7. Flanagan DP, Fiorello CA, Ortiz SO (2010) Enhancing practice through application of Cattell–Horn–Carroll theory and research: a “third method” approach to specific learning disability identification. Psychol Sch 47:739–760. doi: 10.1002/pits Google Scholar
  8. Holland PW (1981) When are item response models consistent with observed data? Psychometrika 46:79–92CrossRefzbMATHMathSciNetGoogle Scholar
  9. Holland PW, Rosenbaum PR (1986) Conditional association and unidimensionality in monotone latent variable models. Ann Stat 14:1523–1543CrossRefzbMATHMathSciNetGoogle Scholar
  10. Junker BW, Ellis JL (1997) A characterization of monotone unidimensional latent variable models. Ann Stat 25:1327–1343CrossRefzbMATHMathSciNetGoogle Scholar
  11. Karlin S, Rinott Y (1980) Classes of orderings of measures and related correlation inequalities, I. Multivariate totally positive distributions. J Multivar Anal 10:467–498CrossRefzbMATHMathSciNetGoogle Scholar
  12. Karlin S, Rinott Y (1983) M-matrices as covariance matrices of multinormal distributions. Linear Algebra Appl 52(53):419–438CrossRefMathSciNetGoogle Scholar
  13. Olsson U (1979) Maximum likelihood estimation of the polychoric correlation coefficient. Psychometrika 44:443–460CrossRefzbMATHMathSciNetGoogle Scholar
  14. Raveh A (1985) On the use of the inverse of the correlation matrix in multivariate data analysis. Am Stat 39:39–42MathSciNetGoogle Scholar
  15. Rinott Y, Scarsini M (2006) Total positivity order and the normal distribution. J Multivar Anal 97:1251–1261CrossRefzbMATHMathSciNetGoogle Scholar
  16. Sardy S, Victoria-Peser M-P (2012) Isotone additive latent variable models. Stat Comput 22:647–659CrossRefMathSciNetGoogle Scholar
  17. Whitt W (1982) Multivariate monotone likelihood ratio and uniform conditional stochastic order. J Appl Probab 19:695–701CrossRefzbMATHMathSciNetGoogle Scholar
  18. Yalcin I, Amemiya Y (2001) Nonlinear factor analysis as a statistical method. Stat Sci 16:275–294CrossRefzbMATHMathSciNetGoogle Scholar
  19. Yung Y-F, Thissen D, McLeod LD (1999) On the relationship between the higher-order factor model and the hierarchical factor model. Psychometrika 64:113–128CrossRefzbMATHMathSciNetGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  1. 1.Radboud University Nijmegen, School of Psychology and Artificial IntelligenceNijmegenThe Netherlands

Personalised recommendations