bi.linking: bi.linking
In eatRest: eatRest

Description Usage Arguments Value Author(s) Examples

Links results from several analysis. Each analysis is linked with each other.

1	bi.linking ( results , folder=NULL , file.name=NULL , method = NULL , lower.triangle = TRUE , scales=NULL )

`results`	result list from automateModels run
`folder`	output folder, will be emptied!
`file.name`	file.name for output excel, default: "bi.linking.results.xlsx"
`method`	set linking method to either "Mean-Mean" , "Haebara" or "Stocking-Lord" (default)
`lower.triangle`	set reference groups for the linking
`scales`	Character vector of scales for which linking should separately done. If NULL, all analysis in the results list are linked. Note: due to suboptimalities in development process, analysis name must contain 'scale'! use this option with care!!

writes linking results to excel file. returns linking results as list.

Martin Hecht

## Not run: 
# 'folder' must be specified, WARNING: this folder is deleted by automateModels!!!
#
# load example data
# this is the results structure returned from running Example 5 of \link{automateModels}
# see there for details of analyses
#
data ( ex5 )
#
# ex5 contains the results of 3 analyses:
names(ex5)
# [1] "all.i__grade.10"  "all.i__grade.9"   "all.i__grade.all"
#
# each pair of these 3 analyses are linked together by bi.linking
# if not run together, you can easily combine analyses from seperate \link{automateModels} runs by calling c(<analysis1>,<analysis2>,...)
# in this case make sure that analysis names are unique
#
# start linking, results are written to folder and are returned 
ex5_linked <- bi.linking ( ex5 , folder = "C:/temp/automateModels/Example5/Linking" , file.name = "ex5_linked" )
#
#
# if you want to link analyses for which no \code{automateModels} results structure is available
# you can use \link{make.link.dummy} to create a structure similar to \code{automateModels} results structure
# this can be used as input for \code{bi.linking}
#
# e.g. let's add an additional analysis for which only item difficulty and standard errors are available
# this information must be in a data.frame, see \link{make.link.dummy} for details
dfr <- data.frame ( 
		"item" = c ( "BioKno01" , "CheKno02" , "PhyKno03" ) ,
		"b"	   = c ( -3.14      , -2.24      , -3.42      ) ,
		"b.se" = c ( 0.612      , 0.453      , 0.783      )
)
#
# create a results object
add <- make.link.dummy ( dfr , "additional" )
#
# add this object to Example 5
ex5add <- c ( add , ex5 )
#
# start the linking procedure with the additional analysis
ex5add_linked <- bi.linking ( ex5add , folder = "C:/temp/automateModels/Example5/Linking2" , file.name = "ex5add_linked" )

## End(Not run)