Description Usage Arguments Details Value Author(s) References See Also Examples
Randomize group membership, and then calculate inter-group phi as
per function phi.inter
.
1 | phi.inter.rnd(iter, grp1, grp2)
|
iter |
Dummy variable; value is ignored |
grp1 |
A data frame, for the first group of participants, that
includes the following columns: |
grp2 |
A data frame, for the second group of participants; format as for grp1. |
Dataset lawson
provides an example of the format expected by
the phi.inter function. For further details, see phi.inter
.
The primary purpose of phi.inter.rnd
is for Monte Carlo
estimation of a chance level for phi.inter
; an example
of usage is given below.
The logic behind this calculation is that if two groups classify the stimulus set differently, their mean phi-inter will be lower than in the case where group membership is determined randomly.
If this is not apparent, consult the brief tutorial available in the Appendix of Lawson et al. (2017).
The phi-inter metric is based on Cramer (1946), slightly developed by Wills & McLaren (1998), and first employed in its current form by Lawson et al. (2017).
Phi-inter; a number ranging between 0 and 1.
Andy J. Wills (andy@willslab.co.uk)
Cramer, H. (1946). Mathematical models of statistics. Princeton, NJ: Princeton University Press.
Lawson, R., Chang, F. & Wills, A.J. (2017). Free classification of large sets of everyday objects is more thematic than taxonomic. Acta Psychologica, 172, 26-40.
Wills, A.J. & McLaren, I.P.L. (1998). Perceptual learning and free classification. Quarterly Journal of Experimental Psychology, 51B, 235-270.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 | ## Estimate phi-inter chance level for the 'lawson' dataset, using 10
## iterations.
data(lawson)
grp1 <- lawson[lawson$ExptGroup == 1,]
grp2 <- lawson[lawson$ExptGroup == 2,]
iterate <- 10
mc <- sapply(1:iterate,phi.inter.rnd,grp1=grp1,grp2=grp2)
mean(mc)
sig.con(0.6,mean(mc))
## In practice, a much larger number of iterations should be used.
## Lawson et al. (2016) use 1e5 iterations.
## At 1e5 iterations this function is very slow. The function has not
## yet been optimised, so peformance may improve in future versions of
## the package.
## If you have a multi-core processor (or Beowulf cluster), speed of
## calculation can be dramatically increased by use of the 'sfSapply'
## function in the 'snowfall' package. Example code follows:
## library(snowfall)
## library(rlecuyer)
## sfInit(parallel=TRUE,cpus=4,type="SOCK")
## sfClusterSetupRNG()
## iterate <- 40
## mc <- sfSapply(1:iterate,phi.inter.rnd,grp1=grp1,grp2=grp2)
## mean(mc)
## sfStop()
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.