-
Notifications
You must be signed in to change notification settings - Fork 0
/
AGRT Python Simulation Study.R
280 lines (255 loc) · 19.5 KB
/
AGRT Python Simulation Study.R
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
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
# To install now archived "mdsdt" package:
install.packages(c("ellipse","polycor"))
install.packages("https://cran.r-project.org/src/contrib/Archive/mdsdt/mdsdt_1.2.tar.gz", type="source", repos=NULL)
# Load required packages
require(mnormt)
require(stringr)
require(mdsdt)
require(scales)
# Generalized subject model
sm.full <- function (x,y,delta,epsilon,sig,m,n,rho, R=NULL) {
# m := 2 element vector specifying the mean spreading for each dimension
# n := 2 element vector specifying the var spreading for each dimension
# rho := 4 element vector specifying the correlation for each quadrant
# sig := 2 element vector specifying the standard deviation for each dimension
# delta := 2 element vector specifying the decision bound intercept for each dimension
# epsilon := 2 element vector specifying the decision bound slopes for each dimension
# calculate decision bounds based on where we are in stimulus space
tau <- delta + epsilon * c(y,x)
# create indicator for which quadrant we're in
zeta <- c(if (x > tau[1]) 1 else -1, if (y > tau[2]) 1 else -1)
# calculate mean vector
mymean <- c(x,y) + rev(zeta) * m * c(x,y)
# calculate sd vector
mysd <- sig + rev(zeta) * n * sig
# pick the rho that corresponds to our quadrant
if (sum(zeta)==2) {
myrho <- rho[4]
} else if (sum(zeta)==-2) {
myrho <- rho[1]
} else if (zeta[1]>0) {
myrho <- rho[3]
} else {
myrho <- rho[2]
}
# create the cov matrix
mycov <- matrix(c(mysd[1]^2, myrho*mysd[1]*mysd[2], myrho*mysd[1]*mysd[2], mysd[2]^2), 2,2,T)
c(
sadmvn(lower=c(-Inf,-Inf), upper=tau, mean=mymean, varcov = mycov)[1],
sadmvn(lower=c(-Inf,tau[2]), upper=c(tau[1],Inf), mean=mymean, varcov = mycov)[1],
sadmvn(lower=c(tau[1],-Inf), upper=c(Inf,tau[2]), mean=mymean, varcov = mycov)[1],
sadmvn(lower=tau, upper=c(Inf,Inf), mean=mymean, varcov = mycov)[1]
)[if (is.null(R)) 1:4 else R]
}
# Subject model used to generate data
test.sm <- function (x,y,correct.resp.ind,sm) {
switch(sm,
"Recovery" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(0, 0), c(0, 0), c(0, 0, 0, 0), R=correct.resp.ind+1),
"PSM" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(.1, 0), c(0, 0), c(0, 0, 0, 0), R=correct.resp.ind+1),
"PSV" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(0, 0), c(0, .4), c(0, 0, 0, 0), R=correct.resp.ind+1),
"PSMV" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(.1, 0), c(0, .4), c(0, 0, 0, 0), R=correct.resp.ind+1),
"PI" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(0, 0), c(0, 0), c(.9, 0, -.9, 0), R=correct.resp.ind+1),
"PI.PSM" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(.1, 0), c(0, 0), c(.9, 0, -.9, 0), R=correct.resp.ind+1),
"PI.PSV" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(0, 0), c(0, .4), c(.9, 0, -.9, 0), R=correct.resp.ind+1),
"PI.PSMV" = .04 * .25 + (1-.04) * sm.full(x, y, c(600, 4), c(0, 0), c(100, .8), c(.1, 0), c(0, .4), c(.9, 0, -.9, 0), R=correct.resp.ind+1))
}
# Import Data
# Without lapse
#mydata <- read.table("/Users/jjg367/Documents/Outside Research/Adaptive GRT/Psi Source/Test_Simulation/data/2023-09-15_21h12.50.080.txt", sep="\t", header=T) # 1 subject, 100 adaptive trials, 1000 GRT trials
#mydata <- read.table("/Users/jjg367/Documents/Outside Research/Adaptive GRT/Psi Source/Test_Simulation/data/2023-09-15_21h13.55.920.txt", sep="\t", header=T) # 100 subjects, 100 adaptive trials, 1000 GRT trials
#mydata <- read.table("/Users/jjg367/Documents/Outside Research/Adaptive GRT/Psi Source/Test_Simulation/data/2023-09-15_23h18.41.285.txt", sep="\t", header=T) # 100 subjects, 300 adaptive trials, 1000 GRT trials
# With lapse
#mydata <- read.table("/Users/jjg367/Documents/Outside Research/Adaptive GRT/Psi Source/Test_Simulation/data/2023-09-17_18h43.35.612.txt", sep="\t", header=T) # 1 subject, 300 adaptive trials, 1000 GRT trials
#mydata <- read.table("/Users/jjg367/Documents/Outside Research/Adaptive GRT/Psi Source/Test_Simulation/data/2023-09-17_18h52.47.370.txt", sep="\t", header=T) # 100 subjects, 300 adaptive trials, 1000 GRT trials
#mydata <- read.table("/Users/jjg367/Documents/Outside Research/Adaptive GRT/Psi Source/Test_Simulation/data/2023-10-02_17h52.55.027.txt", sep="\t", header=T) # 100 subjects, 300 adaptive trials, 1000 GRT trials
# All Subject Models
mydata <- read.table("/Users/jjg367/Documents/Outside Research/Adaptive GRT/Psi Source/Test_Simulation/data/All Models 300 Adaptive Trials 2023-10-02_18h31.13.453.txt", sep="\t", header=T) # 100 subjects, 300 adaptive trials, 1000 GRT trials
# Add Missing Columns
mydata$Dim1 <- sapply(mydata$Stimulus, function (x) as.numeric(unlist(strsplit(str_sub(x,2,-2), ","))[1]))
mydata$Dim2 <- sapply(mydata$Stimulus, function (x) as.numeric(unlist(strsplit(str_sub(x,2,-2), ","))[2]))
mydata$StimInd <- NA
for (sm in unique(mydata$Version)) {
for (s in unique(mydata$Subject[mydata$Version==sm])) {
# 00 01 10 11
d1 <- range(mydata$Dim1[mydata$Block=="main" & mydata$Subject==s & mydata$Version==sm])
d2 <- range(mydata$Dim2[mydata$Block=="main" & mydata$Subject==s & mydata$Version==sm])
mydata$StimInd[mydata$Block=="main" & mydata$Subject==s & mydata$Version==sm] <- sapply(mydata$Stimulus[mydata$Block=="main" & mydata$Subject==s & mydata$Version==sm], function (x) {
mystim <- eval(parse(text=paste0("c",x)))
if (mystim[1] == d1[1] & mystim[2] == d2[1]) 0 else
if (mystim[1] == d1[1] & mystim[2] == d2[2]) 1 else
if (mystim[1] == d1[2] & mystim[2] == d2[1]) 2 else
if (mystim[1] == d1[2] & mystim[2] == d2[2]) 3
})
}
}
mydata$RespInd <- NA; mydata$RespInd[mydata$Response=="(0, 0)"] <- 0; mydata$RespInd[mydata$Response=="(0, 1)"] <- 1; mydata$RespInd[mydata$Response=="(1, 0)"] <- 2; mydata$RespInd[mydata$Response=="(1, 1)"] <- 3
mydata$Correct <- mydata$StimInd == mydata$RespInd
mydata$Alpha1 <- sapply(mydata$Lambda, function (x) as.numeric(unlist(strsplit(gsub("[()]","",x), ",")))[1])
mydata$Beta1 <- sapply(mydata$Lambda, function (x) as.numeric(unlist(strsplit(gsub("[()]","",x), ",")))[2])
mydata$Alpha2 <- sapply(mydata$Lambda, function (x) as.numeric(unlist(strsplit(gsub("[()]","",x), ",")))[3])
mydata$Beta2 <- sapply(mydata$Lambda, function (x) as.numeric(unlist(strsplit(gsub("[()]","",x), ",")))[4])
mydata$Pred.Dim1min <- mydata$Alpha1 - mydata$Beta1 * sqrt(2) * erfinv((2*sqrt(.75)-.04)/(1-.04)-1)
mydata$Pred.Dim1max <- mydata$Alpha1 - mydata$Beta1 * sqrt(2) * erfinv((2*(1-sqrt(.75))-.04)/(1-.04)-1)
mydata$Pred.Dim2min <- mydata$Alpha2 - mydata$Beta2 * sqrt(2) * erfinv((2*sqrt(.75)-.04)/(1-.04)-1)
mydata$Pred.Dim2max <- mydata$Alpha2 - mydata$Beta2 * sqrt(2) * erfinv((2*(1-sqrt(.75))-.04)/(1-.04)-1)
mydata$Pred.Acc <- NA
mydata$Pred.Acc[mydata$Block=="adapt"] <- mapply(function (x1,x2,y1,y2,z) {
mean(c(test.sm(x1,y1,0,z), test.sm(x1,y2,1,z), test.sm(x2,y1,2,z), test.sm(x2,y2,3,z)))
}, x1=mydata$Pred.Dim1min[mydata$Block=="adapt"], x2=mydata$Pred.Dim1max[mydata$Block=="adapt"],
y1=mydata$Pred.Dim2min[mydata$Block=="adapt"], y2=mydata$Pred.Dim2max[mydata$Block=="adapt"], z = mydata$Version[mydata$Block=="adapt"])
# For only one subject:
# par(mfrow=c(2,2))
# plot(mydata$Trial[mydata$Block=="adapt"], mydata$Alpha1[mydata$Block=="adapt"], type='l', main="Dimension 1", xlab="Trial", ylab="Alpha (Decision Threshold)"); abline(h=600,lty=2)
# plot(mydata$Trial[mydata$Block=="adapt"], mydata$Beta1[mydata$Block=="adapt"], type='l', main="Dimension 1", xlab="Trial", ylab="Beta (Stimulus Separation)"); abline(h=100,lty=2)
# plot(mydata$Trial[mydata$Block=="adapt"], mydata$Alpha2[mydata$Block=="adapt"], type='l', main="Dimension 2", xlab="Trial", ylab="Alpha (Decision Threshold)"); abline(h=4,lty=2)
# plot(mydata$Trial[mydata$Block=="adapt"], mydata$Beta2[mydata$Block=="adapt"], type='l', main="Dimension 2", xlab="Trial", ylab="Beta (Stimulus Separation)"); abline(h=.8,lty=2)
# plot(mydata$Trial[mydata$Block=="adapt"], mydata$Pred.Acc[mydata$Block=="adapt"], type='l', main="Predicted Accuracy", xlab="Trial", ylab="Overall Accuracy"); abline(h=.75, lty=2)
# par(mfrow=c(1,1))
# SINGLE SUBJECT MODEL, MULTIPLE SUBJECTS
# Plot Psi Parameters
# par(mfrow=c(2,2))
# plot(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Alpha1[mydata$Block=="adapt" & mydata$Trial==x])
# }), type='l', main="Dimension 1", xlab="Trial", ylab="Alpha (Decision Threshold)", ylim=c(100,800)); abline(h=600,lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Alpha1[mydata$Block=="adapt" & mydata$Trial==x]) + sd(mydata$Alpha1[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Alpha1[mydata$Block=="adapt" & mydata$Trial==x]) - sd(mydata$Alpha1[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# plot(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Beta1[mydata$Block=="adapt" & mydata$Trial==x])
# }), type='l', main="Dimension 1", xlab="Trial", ylab="Beta (Stimulus Separation)", ylim=c(0,200)); abline(h=100,lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Beta1[mydata$Block=="adapt" & mydata$Trial==x]) + sd(mydata$Beta1[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Beta1[mydata$Block=="adapt" & mydata$Trial==x]) - sd(mydata$Beta1[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# plot(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Alpha2[mydata$Block=="adapt" & mydata$Trial==x])
# }), type='l', main="Dimension 2", xlab="Trial", ylab="Alpha (Decision Threshold)", ylim=c(1,12)); abline(h=4,lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Alpha2[mydata$Block=="adapt" & mydata$Trial==x]) + sd(mydata$Alpha2[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Alpha2[mydata$Block=="adapt" & mydata$Trial==x]) - sd(mydata$Alpha2[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# plot(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Beta2[mydata$Block=="adapt" & mydata$Trial==x])
# }), type='l', main="Dimension 2", xlab="Trial", ylab="Beta (Stimulus Separation)", ylim=c(0,2)); abline(h=.8,lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Beta2[mydata$Block=="adapt" & mydata$Trial==x]) + sd(mydata$Beta2[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Beta2[mydata$Block=="adapt" & mydata$Trial==x]) - sd(mydata$Beta2[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# par(mfrow=c(1,1))
# # Plot Predicted Accuracy
# plot(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Pred.Acc[mydata$Block=="adapt" & mydata$Trial==x])
# }), type='l', main="Predicted Accuracy", xlab="Trial", ylab="Overall Accuracy", ylim=c(0,1)); abline(h=.75, lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Pred.Acc[mydata$Block=="adapt" & mydata$Trial==x]) + sd(mydata$Pred.Acc[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# lines(unique(mydata$Trial[mydata$Block=="adapt"]), sapply(unique(mydata$Trial[mydata$Block=="adapt"]), function (x) {
# mean(mydata$Pred.Acc[mydata$Block=="adapt" & mydata$Trial==x]) - sd(mydata$Pred.Acc[mydata$Block=="adapt" & mydata$Trial==x])
# }), lty=2)
# # Plot Actual Accuracy
# barplot(sapply(sort(unique(mydata$Subject)), function (x) mean(mydata$Correct[mydata$Block=="main" & mydata$Subject==x])),
# main="Overall Accuracy", xlab="Simulated Subject", names.arg=sort(unique(mydata$Subject)), ylab="Accuracy", ylim=c(0,1)); abline(h=.75, lty=2)
# # Plot fitted GRT models for first 10 subjects
# for (s in 0:10) plot(fit.grt(table(mydata$StimInd[mydata$Block=="main" & mydata$Subject==s],mydata$RespInd[mydata$Block=="main" & mydata$Subject==s], exclude=NA)), main=s, marginals=T)
# ALL SUBJECT MODELS ALL SUBJECTS
# Use Shading
# x loc
# PSM MODELS IN RED
plot(0,0,type='n',xlim=c(0,300),ylim=c(400,800),xlab="Trial",ylab="Parameter Estimate",main="X Threshold")
abline(h=600,lty=1, lwd=2, xpd=F)
for (sm in unique(mydata$Version)) {
conv.data <- subset(mydata, Block=="adapt" & Version==sm)
lines(sort(unique(conv.data$Trial)), sapply(sort(unique(conv.data$Trial)), function (x) mean(conv.data$Alpha1[conv.data$Trial==x])),
col=if (sm %in% c("PSM","PSMV","PI.PSM","PI.PSMV")) 'red' else 'black')
polygon(c(sort(unique(conv.data$Trial)),rev(sort(unique(conv.data$Trial)))),
c(sapply(sort(unique(conv.data$Trial)),
function (x) mean(conv.data$Alpha1[conv.data$Trial==x]) + sd(conv.data$Alpha1[conv.data$Trial==x])),
sapply(rev(sort(unique(conv.data$Trial))),
function (x) mean(conv.data$Alpha1[conv.data$Trial==x]) - sd(conv.data$Alpha1[conv.data$Trial==x]))),
col=alpha(if (sm %in% c("PSM","PSMV","PI.PSM","PI.PSMV")) 'red' else 'black', alpha=.05), border=NA)
}
legend('top', legend=c("Violate Perceptual Separability", "Other"), fill=c('red','black'), horiz=T, bty='n', cex=.6)
# x slope
plot(0,0,type='n',xlim=c(0,300),ylim=c(0,200),xlab="Trial",ylab="Parameter Estimate",main="X SD")
abline(h=100,lty=1,lwd=2, xpd=F)
for (sm in unique(mydata$Version)) {
conv.data <- subset(mydata, Block=="adapt" & Version==sm)
lines(sort(unique(conv.data$Trial)), sapply(sort(unique(conv.data$Trial)), function (x) mean(conv.data$Beta1[conv.data$Trial==x])),
col=if (sm %in% c("PSM","PSMV","PI.PSM","PI.PSMV")) 'red' else 'black')
polygon(c(sort(unique(conv.data$Trial)),rev(sort(unique(conv.data$Trial)))),
c(sapply(sort(unique(conv.data$Trial)),
function (x) mean(conv.data$Beta1[conv.data$Trial==x]) + sd(conv.data$Beta1[conv.data$Trial==x])),
sapply(rev(sort(unique(conv.data$Trial))),
function (x) mean(conv.data$Beta1[conv.data$Trial==x]) - sd(conv.data$Beta1[conv.data$Trial==x]))),
col=alpha(if (sm %in% c("PSM","PSMV","PI.PSM","PI.PSMV")) 'red' else 'black', alpha=.05), border=NA)
}
legend('top', legend=c("Violate Perceptual Separability", "Other"), fill=c('red','black'), horiz=T, bty='n', cex=.6)
# y loc
plot(0,0,type='n',xlim=c(0,300),ylim=c(0,8),xlab="Trial",ylab="Parameter Estimate",main="Y Threshold")
abline(h=4,lty=1,lwd=2,xpd=F)
for (sm in unique(mydata$Version)) {
conv.data <- subset(mydata, Block=="adapt" & Version==sm)
lines(sort(unique(conv.data$Trial)), sapply(sort(unique(conv.data$Trial)), function (x) mean(conv.data$Alpha2[conv.data$Trial==x])),
col=if (sm %in% c()) 'red' else 'black')
polygon(c(sort(unique(conv.data$Trial)),rev(sort(unique(conv.data$Trial)))),
c(sapply(sort(unique(conv.data$Trial)),
function (x) mean(conv.data$Alpha2[conv.data$Trial==x]) + sd(conv.data$Alpha2[conv.data$Trial==x])),
sapply(rev(sort(unique(conv.data$Trial))),
function (x) mean(conv.data$Alpha2[conv.data$Trial==x]) - sd(conv.data$Alpha2[conv.data$Trial==x]))),
col=alpha(if (sm %in% c()) 'red' else 'black', alpha=.05), border=NA)
}
# y slope
plot(0,0,type='n',xlim=c(0,300),ylim=c(0,2),xlab="Trial",ylab="Parameter Estimate",main="Y SD")
abline(h=.8,lty=1,lwd=2, xpd=F)
for (sm in unique(mydata$Version)) {
conv.data <- subset(mydata, Block=="adapt" & Version==sm)
lines(sort(unique(conv.data$Trial)), sapply(sort(unique(conv.data$Trial)), function (x) mean(conv.data$Beta2[conv.data$Trial==x])),
col=if (sm %in% c()) 'red' else 'black')
polygon(c(sort(unique(conv.data$Trial)),rev(sort(unique(conv.data$Trial)))),
c(sapply(sort(unique(conv.data$Trial)),
function (x) mean(conv.data$Beta2[conv.data$Trial==x]) + sd(conv.data$Beta2[conv.data$Trial==x])),
sapply(rev(sort(unique(conv.data$Trial))),
function (x) mean(conv.data$Beta2[conv.data$Trial==x]) - sd(conv.data$Beta2[conv.data$Trial==x]))),
col=alpha(if (sm %in% c()) 'red' else 'black', alpha=.05), border=NA)
}
# acc
# PI MODELS IN RED
plot(0,0,type='n',xlim=c(0,300),ylim=c(.5,1),xlab="Trial",ylab="Accuracy Estimate")#,main="Predicted Accuracy")
abline(h=.75,lty=1,lwd=2, xpd=F)
for (sm in unique(mydata$Version)) {
conv.data <- subset(mydata, Block=="adapt" & Version==sm)
lines(sort(unique(conv.data$Trial)), sapply(sort(unique(conv.data$Trial)), function (x) mean(conv.data$Pred.Acc[conv.data$Trial==x])),
col=if (sm %in% c("PI","PI.PSM","PI.PSV","PI.PSMV")) 'red' else 'black')
polygon(c(sort(unique(conv.data$Trial)),rev(sort(unique(conv.data$Trial)))),
c(sapply(sort(unique(conv.data$Trial)),
function (x) mean(conv.data$Pred.Acc[conv.data$Trial==x]) + sd(conv.data$Pred.Acc[conv.data$Trial==x])),
sapply(rev(sort(unique(conv.data$Trial))),
function (x) mean(conv.data$Pred.Acc[conv.data$Trial==x]) - sd(conv.data$Pred.Acc[conv.data$Trial==x]))),
col=alpha(if (sm %in% c("PI","PI.PSM","PI.PSV","PI.PSMV")) 'red' else 'black', alpha=.05), border=NA)
}
legend('top', legend=c("Violate Perceptual Independence", "Other"), fill=c('red','black'), horiz=T, bty='n', cex=.6)
# Calculate assumption violation rates
grtResults <- matrix(nrow=2, ncol=32); m <- 0
for (sm in c("Recovery","PSM","PSV","PSMV","PI","PI.PSM","PI.PSV","PI.PSMV")) {
m <- m + 1
this.data <- subset(mydata, Version == sm)
violations <- lapply(sort(unique(this.data$Subject)), function (s) c(1 * (riTest(table(this.data$StimInd[this.data$Block=="main" & this.data$Subject==s],this.data$RespInd[this.data$Block=="main" & this.data$Subject==s], exclude=NA))$p.value < .05),
1 * (mriTest(table(this.data$StimInd[this.data$Block=="main" & this.data$Subject==s],this.data$RespInd[this.data$Block=="main" & this.data$Subject==s], exclude=NA))$p.value < .05)))
violations <- sapply(1:8, function (x) sum(sapply(1:length(unique(this.data$Subject)), function (y) violations[[y]][x]), na.rm=T)/sum(sapply(1:length(unique(this.data$Subject)), function (y) !is.na(violations[[y]][x]))))
grtResults[1, 1:4 + 4*(m-1)] <- violations[1:4]
grtResults[2, 1:4 + 4*(m-1)] <- violations[5:8]
}
barplot(grtResults, beside=T, main="", xlab="Stimulus/Response", ylab=paste("Null Hypothesis Rejection Rate (",expression(alpha)," = .05)",sep=''),
legend.text = c("Report Independence", "Marginal Response Invariance"),
args.legend = list(title="Test", bty='n', horiz=T, xpd=NA, x=48,y=-.2, xjust=.5,cex=.8, col='gray'))
abline(v=seq(12.5,by=12,length.out = length(unique(mydata$Version))-1), lty=2)
# Plot stimulus intensity estimates