This function will test the 'anydifferences' hypothesis (conceptually a oneway ANOVA test for dprimes) with one of the Wald, Pearson or likelihood ratio chisquare test statistics. The common dprime is estimated with ML or weighted average.
1 2 3  dprime_compare(correct, total, protocol, conf.level = 0.95,
statistic = c("likelihood", "Pearson", "Wald.p", "Wald.d"),
estim = c("ML", "weighted.avg"))

correct 
a numeric vector of the number of correct answers; one element for each test. 
total 
a numeric vector of the total number of trials; one element for each test. 
protocol 
a character vector or factor naming the protocol used; one element
for each test. Currently the following protocols are supported:

conf.level 
the confidence level for the estimated common dprime. 
statistic 
the test statistic for testing the 'anydifferences' hypothesis. 
estim 
The estimation method for the common dprime. 
The vectors correct
, total
and protocol
have to
be of the same length.
The function has a print method.
an object of class "dprime_compare"
with the following elements
stat.value 
the value of the (chisquare) test statistic for the 'anydifferences' hypothesis. 
df 
the degrees of freedom for the 
p.value 
the pvalue for the 'anydifferences' test. 
statistic 
the name of the test statistic for the 'anydifferences' test. 
data 
the data table produced by 
coefficients 
'table' with estimated common dprime, standard error and confidence
limits storred as a onerow 
conf.level 
confidence level for the common dprime. 
conf.int 
the confidence interval for the common dprime. 
estim 
the estimation method for the common dprime. 
conf.method 
the statistical method/test statistic used to compute the confidence interval for the common dprime. 
Rune Haubo B Christensen
dprime_test
, dprime_table
,
posthoc.dprime_compare
.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15  ## Make some fake data:
n < rep(40, 4)
x < c(25, 25, 30, 35)
protocol < c("triangle", "duotrio", "threeAFC", "twoAFC")
## Look at the data table with dprimes etc.:
dprime_table(x, n, protocol)
## 'any differences' test:
## ML estimation and test with likelihood statistic:
(dpc < dprime_compare(x, n, protocol))
## Other estimation/statistic options:
dprime_compare(x, n, protocol, estim="weighted.avg")
dprime_compare(x, n, protocol, statistic="Pearson")
dprime_compare(x, n, protocol, statistic="Wald.p")
dprime_compare(x, n, protocol, statistic="Wald.d")

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.
Please suggest features or report bugs with the GitHub issue tracker.
All documentation is copyright its authors; we didn't write any of that.