MLDS (version 0.1-5)

simu.6pt: Perform Bootstrap Test on 6-point Likelihood for MLDS FIT

Description

Using the fitted responses (probabilities) to the difference scale, new responses are generated which permit new 6-point likelihoods to be calculated. The distribution of a large number of such likelihoods can be compared with that obtained from the observed responses to evaluate the internal consistency of the estimated scale.

Usage

simu.6pt(obj, nsim = 1)

Arguments

obj

object of class mlds

nsim

integer indicating number of bootstrap trials.

Value

• LIST with 4 components
• boot.samp
  vector of numeric giving the log likelihood for the 6-point test for each simulation.
• lik6pt
  numeric indicating the log likelihood for the 6-point test on the original data
• p
  proportion of simulations on which the simulated log likelihood was higher than that obtained from the original sample.
• N
  numeric indicating the number of simulations. It should be the length of boot.samp.

References

Maloney, L. T. and Yang, J. N. (2003). Maximum likelihood difference scaling. Journal of Vision, 3(8):5, 573--585, http://journalofvision.org/3/8/5/, doi:10.1167/3.8.5.

See Also

mlds, lik6pt

Examples

data(kk1)
	x.mlds <- mlds(SwapOrder(kk1))
	#nsim should be near 10,000 for stability,
	# but this will take a little time
	simu.6pt(x.mlds, 100)