Both the 3PLM and the 3PNO define item characteristic curves that are identical
up to a scaling constant d = 1.702 . To estimate the item and ability parameters in this model a Gibbs sampler (GEMAN; GEMAN, 1984) might be applied. Among others, Albert (1992), Béguin e Glas (2001), Fox e Glas (2001), and Matteucci, Mignane e Veldkamp (2009) developed Gibbs samplers for various IRT models. Standardized software packages, like MATLAB or the R-package, can be applied for Bayesian parameter estimation. Albert (1992) provides MATLAB code for the 2PNO model and Fox (2010) provides R codes for several more complicated IRT models.