check.plot.mcmc.3pnob {bairt} | R Documentation |
Marginal Posterior means of b_j plotted against the marginal posterior means of a_j. Each point is labeled with the number of the corresponding Item.
For the Three-Parameter Normal Ogive Item Response Model (3pno), the size of the numbers refers to the marginal posterior means of c_j.
The Potential Scale Reduction Factor (Rhat) is calculated for each chain, bairt generates a single MCMC and evaluates convergence by breaking the chain in three sub chains and comparing the between- and within-subchain variance.
The black color suggests convergence and red items indicate convergence problems (Rhat greater than 1.1).
## S3 method for class 'mcmc.3pnob' check.plot(mcmclist, converg.test = T, c.probs = c(0, 0.2, 0.5, 1), legen = "topleft", ...)
mcmclist |
A mcmc.2pnob or mcmc.3pnob class object. |
converg.test |
Checking if Rhat is major that 1.1. |
c.probs |
Vector for assignment of intervals the Guessing (c). |
legen |
Coordinates to be used to position the Guessing (c) legend. |
... |
Further arguments. |
If converg.test = TRUE the items with Rhat menor that 1.1 are print in red color. It is useful for quick check of the convergence.
A plot of the discrimination marginal posterior means against difficulty marginal posterior means. For the Three-parameter model the guessing marginal posterior means are represented by the number size of the item.
Javier MartÃnez
Johnson, V. E. & Albert, J. H. (1999). Ordinal Data Modeling. New York: Springer.
Gelman, A., Carlin, J. B., Stern, H. S. & Rubin, B. (2004). Bayesian Data Analysis.New York: Chapman & Hall/CRC.
mcmc.2pnob
, mcmc.3pnob
and
continue.mcmc.bairt
.
# data for model data("MathTest") # Only for the first 500 examinees of the data MathTest # Two-Parameter Normal Ogive Model model2 <- mcmc.2pnob(MathTest[1:500,], iter = 400, burning = 100) check.plot(model2) chain.study(model2, parameter = "b", chain = 12) chain.study(model2, parameter = "theta", chain = 10) # For all examinees of the data # Two-Parameter Normal Ogive Model modelAll2 <- mcmc.2pnob(MathTest, iter = 3500, burning = 500, thin = 10) check.plot(modelAll2) chain.study(modelAll2, parameter = "b", chain = 14) chain.study(modelAll2, parameter = "theta", chain = 10) # Three-Parameter Normal Ogive Model modelAll3 <- mcmc.3pnob(MathTest, iter = 3500, burning = 500, thin = 10) check.plot(modelAll3) chain.study(modelAll3, parameter = "b", chain = 12) chain.study(modelAll3, parameter = "c", chain = 10) ## End(Not run)