MPDiR (version 0.2)

psyfun.boot: Bootstrapping Standard Errors of Psychometric Function Parameters

Description

A function that will run a bootstrap on the estimated parameters of a psychometric function fit given a model object.

Usage

psyfun.boot(obj, N = 100)

Value

Returns a matrix with one row for each coefficient of the model and one column for each bootstrap replication.

Arguments

obj

object inheriting from class ‘glm’ from a fit of a psychometric function

N

integer indicating number of bootstrap replications.

Author

Kenneth Knoblauch

Details

The function computes new binomial responses based on the fitted probabilities of the model object for each bootstrap replication. A psychometric function is then fit to each one and the fitted coefficients returned as a bootstrap replicate.

References

Maloney, L. T. (1990) Confidence interval for the parameters of psychometric functions. Perception & Psychophysics, 47(2), 127--134.

Foster, D.H., Bischof, W.F.(1997) Bootstrap estimates of the statistical accuracy of thresholds obtained from psychometric functions. Spatial Vision, 11(1), 135--139.

Treutwein, B., Strasburger, H. (1999) Fitting the psychometric function. Perception & Psychophysics, 61(1), 87--106.

Examples

data(HSP)
SHR2 <- subset(HSP, Obs == "SH" & Run == "R2")
SHR2 <- within(SHR2, {
	nyes <- N * p/100
	nno <- N - nyes
	})
SHR2.glm <- glm(cbind(nyes, nno) ~ log(Q), binomial, SHR2)
### For a real problem, set N to 10000 or so
SHR2.boot <- psyfun.boot(SHR2.glm, 10)