thresholdssetter {QCA3} | R Documentation |
If the conditions is continuous variable, it has to be recoded into
binary variable in order to conduct csQCA, or multi-value variable in
order to conduct mvQCA. thresholdsster
helps to recode the
variable using cluster analysis (the default method is 'average linkage').
thresholdssetter(x, nthreshold = 1, value = TRUE, method ="average", thresholds=NULL,dismethod="euclidean",print.table=TRUE)
x |
A continouse variable. |
nthreshold |
The number of thresholds. |
value |
logical, returns the recoded variable when TRUE, otherwise returns the thresholods only. |
method |
method of cluster analysis. See hclust for more
details. |
thresholds |
numeric vector of the threshold values. If it is not NULL, then recode the variable according to thresholds rather than cluster analysis. The threshold values are included in the smaller group. |
dismethod |
method argument for dist . |
print.table |
logical, if TRUE, print the table of the recoded variable. |
In order to dichotomize the data, 1 threshold is needed; to trichotomize the data, 2 thresholds are needed; and so on. Yet, it should not be a large number, otherwise, limited diversity is a problem. It is suggested to be 1 to 3.
Whenever possible, use theoretically meaningful thresholds. Use statistical means only when necessary. The thresholds should make theoretical sense, thus you should examine the threshold values before going on. Good thresholod values should NOT create very differently sized subgroups. Mean and median should be avoided as well.
A vector of the thresholds when value is FALSE. The recoded variable when value is TRUE.
Ronggui HUANG
Cronqvist, Lasse and Berg-Schlosser, Dirk. 2009. Multi-Value QCA (mvQCA). In Configuraional comparative Methods: qualitative comparative analysis (QCA) and related techniques. ed by Benoit RiHoux and Charles Ragin. Sage.
Cronqvist, L. 2007. Tosmana user manual. http://www.tosmana.net/tosmana_manual1_3beta.pdf
cuttingPoint <- thresholdssetter(Lipset$GNPCAP,2,value=FALSE) thresholdssetter(Lipset$GNPCAP,thresholds=cuttingPoint) thresholdssetter(Lipset$GNPCAP,2)## the same as the previous one thresholdssetter(Lipset$GNPCAP,2,print=FALSE) ## print recoded variable rather than a table