Maximum Likelihood Difference Scaling

Difference scaling is a method for scaling perceived supra-threshold differences. The package contains functions that allow the user to design and run a difference scaling experiment, to fit the resulting data by maximum likelihood and test the internal validity of the estimated scale.


News

version 0.1-2

made mlds generic and added default (for glm and optim) and formula methods. The formula method uses optim.

version 0.1-3

added function df2mlds.df to coerce data frame from results to object of class mlds.df with appropriate attributes

version 0.1-5

modified default DisplaySize in runSampleExperiment

Changed argument N to PntNum in DisplayOneTrial to avoid name conflicts due to partial matching

version 0.1-6

modified LazyData yes

added aspect argument to runSampleExperiment for controlling aspect ratio of graphic display

graphic display in runSampleExperiment now opens using dev.new for better platform independence

added xlim and ylim arguments to DisplayOneTrial for controlling size of displayed scatterplots

fixed some unmatched braces in Rd files

version 0.1-7

added as.mlds.df S3 generic

changed df2mlds.df and ix.mat2df to as.mlds.df.data.frame and deprecated both of the former functions.

version 0.1-8

modified boot.mlds to use glm.fit so about 2x faster

fixed unmatched parenthesis in mlds.Rd

fixed bug in runSampleExperiment in which program bombs if observer enters a carriage return instead of a number

made runSampleExperiment not bomb if observer enters arbitrary character

version 0.2-0

fixed warnings in mlds.Rd

eliminated AIC method, relying on default (suggested by Achim Zeileis)

fixed some non ASCII text in Rd files

added coef and vcov methods for object of class mlds

version 0.2-1

Fixed some Rd files that were giving warnings

version 0.2-2

Method of triads integrated into package. mlds.default has been removed and now there is mlds.mlds.df for the method of quadruples (input objects of class mlds.df or data.frame) and mlds.mlbs.df for the method of triads (objects of class mlbs.df or data.frame) and generating objects of class mlds and mlbs, respectively.

RunSampleExperiment changed to RunQuadExperiment and RunTriadExperiment added. The first produces an object of class mlds.df, as the older function did and the latter an object of class mlbs.df

complementary methods added for objects of class mlbs.df and class mlbs

mlds.formula method works for mlbs.df objects, i.e., triads, too.

version 0.2-3

Get6pts wasn't taking into account multiple trials with the same quadruples. Fixed now.

Get6pts and simu.6pt take a new argument, nrep, which indicates how many sessions (complete sets of quadruples) the data represent.

version 0.2-4

added CITATION file to inst directory

version 0.2-5

fixed bug in runQuadExperiment example on man page

version 0.2-6

typos in Rd files

re-ordered ... arguments to optim in mlds.formula method

version 0.2-7

added function SimMLDS which creates data from a simulated MLDS experiment

version 0.2-8

suppressed warnings from boot.mlds, boot.mlbs binom.diagnostics, and simu.6pt and added option no.warn to each of these to turn them back on

made the output of boot.mlds and boot.mlbs lists of class "mlds.bt" and added a summary method to extract the means and SDs. These can be specified to be either on the standard or unnormalized scales

added defunct function for runSampleExperiment, this being replaced earlier by runQuadExperiment and runTriadExperiment

version 0.2-9

added vignette

added Transparency data set

version 0.3-1

added glm.meth argument to mlds.mlds.df and mlds.mlbs.df so that other methods could be passed to glm function, such a brglm.fit from the brglm package.

version 0.3-2

removed LazyLoad from Description

moved vignette from inst/doc to vignettes

version 0-4.0

mlds.mlds.df and mlds.mlbs.df changed so that model matrices constructed
internally by formula objects and the model.matrix function.

version 0-4.1

fixed bug in predict.mlds and predict.mlbs when using newdata argument. mlds now works for triads when only a subset of the data frame is used.

version 0-4.2

added argument to plot, lines, points methods to plot difference scales in units of d'.

removed some unnecessary triple colons in referencing packages

version 0-4.4

fixed formatting error in vignette

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("MLDS")

0.4.5 by Ken Knoblauch, 3 years ago


Browse source code at https://github.com/cran/MLDS


Authors: Kenneth Knoblauch and Laurence T. Maloney , based in part on C code written by Laurence T. Maloney and J. N. Yang


Documentation:   PDF Manual  


Task views: Psychometric Models and Methods


GPL (>= 2) license


Imports MASS

Depends on graphics, stats, utils, base


See at CRAN