-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy path[6].Sensitivity_analyses.R
259 lines (174 loc) · 8.91 KB
/
[6].Sensitivity_analyses.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
# Martin R. Vasilev, 2017
rm(list=ls())
source("Functions/settings.R")
library(ggplot2)
#------------------------------------------------
# Plot of how results differ based on the priors:
#------------------------------------------------
# load up summary files:
load("Summary/G1s.Rda") # general sound effect, reading comprehension (uniform priors)
load("Summary/G2s.Rda") # general sound effect, reading comprehension (normal priors)
load("Summary/GR1s.Rda") # general sound effect, reading speed (uniform priors)
load("Summary/GR2s.Rda") # general sound effect, reading speed (normal priors)
load("Summary/M1s.Rda") # music effect, reading comprehension (uniform priors)
load("Summary/M2s.Rda") # music effect, reading comprehension (normal priors)
load("Summary/N1s.Rda") # noise effect, reading comprehension (uniform priors)
load("Summary/N2s.Rda") # noise effect, reading comprehension (normal priors)
load("Summary/S1s.Rda") # speech effect, reading comprehension (uniform priors)
load("Summary/S2s.Rda") # speech effect, reading comprehension (normal priors)
muMean<- c(G1s$statistics[1,1], G2s$statistics[1,1], GR1s$statistics[1,1],
GR2s$statistics[1,1], M1s$statistics[1,1], M2s$statistics[1,1],
N1s$statistics[1,1], N2s$statistics[1,1], S1s$statistics[1,1],
S2s$statistics[1,1])
tauMean<- c(G1s$statistics[2,1], G2s$statistics[2,1], GR1s$statistics[2,1],
GR2s$statistics[2,1], M1s$statistics[2,1], M2s$statistics[2,1],
N1s$statistics[2,1], N2s$statistics[2,1], S1s$statistics[2,1],
S2s$statistics[2,1])
Model<- c(rep("All sounds (compr.)",2), rep("All sounds (read. speed)",2),
rep("Music (compr.)",2), rep("Noise (compr.)",2), rep("Speech (compr.)",2))
p1<- expression(paste(theta, "~ Unif(-10, 10); ", tau, "~ Unif(0, 10)"))
p2<- expression(paste(theta, "~ Normal(0, ", 100^{2}, "); ", tau, "~ Normal(0, ", 100^{2}, ") I(0,)"))
Priors<- c(rep(c("Uniform", "Normal"),5))
MuCrIl<- c(G1s$quantiles[1,1], G2s$quantiles[1,1], GR1s$quantiles[1,1],
GR2s$quantiles[1,1], M1s$quantiles[1,1], M2s$quantiles[1,1],
N1s$quantiles[1,1], N2s$quantiles[1,1], S1s$quantiles[1,1],
S2s$quantiles[1,1])
MuCrIr<- c(G1s$quantiles[1,5], G2s$quantiles[1,5], GR1s$quantiles[1,5],
GR2s$quantiles[1,5], M1s$quantiles[1,5], M2s$quantiles[1,5],
N1s$quantiles[1,5], N2s$quantiles[1,5], S1s$quantiles[1,5],
S2s$quantiles[1,5])
TauCrIl<- c(G1s$quantiles[2,1], G2s$quantiles[2,1], GR1s$quantiles[2,1],
GR2s$quantiles[2,1], M1s$quantiles[2,1], M2s$quantiles[2,1],
N1s$quantiles[2,1], N2s$quantiles[2,1], S1s$quantiles[2,1],
S2s$quantiles[2,1])
TauCrIr<- c(G1s$quantiles[2,5], G2s$quantiles[2,5], GR1s$quantiles[2,5],
GR2s$quantiles[2,5], M1s$quantiles[2,5], M2s$quantiles[2,5],
N1s$quantiles[2,5], N2s$quantiles[2,5], S1s$quantiles[2,5],
S2s$quantiles[2,5])
#db<- data.frame(muMean, tauMean, Model, Priors)
xpos<- c(0,3,6,9,12)
png('Plots/Prior_sensitivity.png', width = 3600, height = 1600, units = "px", res=300, type="cairo")
#par(mfrow=c(1,2), mar= c(5,5,5,1))
layout(mat = matrix(c(1,1,2,3),nrow = 2,ncol = 2,byrow = TRUE),heights = c(0.35,0.75))
par(mar=c(4,4.3,0,3))
# legend:
plot(0, type = "n", axes=FALSE, xlab="", ylab="")
legend(x = "bottom",inset = 0,
legend = c(p1, p2),
col=c("darkred", "darkorange"), lwd=3, cex=1.3, horiz = FALSE,
title = expression(bold("Priors")))
plot(NA,ylab='Effect size (g)',xlab='Model', ylim= c(round(min(MuCrIl),2)-0.1,
round(max(MuCrIr),2))+0.05,
xlim= c(-1, 13), xaxt='n',
family="serif", cex.axis=1.4, cex.lab=1.4,xaxs='i', yaxs='i')
#par(xpd=TRUE)
axis(side = 1, at = xpos, tick = T, labels = c("All (RC)", "All (RS)",
"Music (RC)", "Noise (RC)", "Speech (RC)"), padj = 0.5, cex=1.6)
#mtext("Model", line=4, side = 1, cex = 1.4, family="serif")
# Uniform means:
points(x= xpos-0.5, y= muMean[c(1,3,5,7,9)], pch = 19, cex=1.2, col="darkred")
# Normal means:
points(x= xpos+0.5, y= muMean[c(2,4,6,8,10)], pch = 15, cex=1.2, col="darkorange")
# Uniform CrIs:
segments(x0=xpos-0.5, y0=MuCrIl[c(1,3,5,7,9)], x1=xpos-0.5, y1=MuCrIr[c(1,3,5,7,9)],
lwd=1.2, col="darkred")
segments(x0=xpos-0.35, y0=MuCrIl[c(1,3,5,7,9)], x1=xpos-0.65, y1=MuCrIl[c(1,3,5,7,9)],
lwd=1.2, col="darkred")
segments(x0=xpos-0.35, y0=MuCrIr[c(1,3,5,7,9)], x1=xpos-0.65, y1=MuCrIr[c(1,3,5,7,9)],
lwd=1.2, col="darkred")
# Normal CrIs:
segments(x0=xpos+0.5, y0=MuCrIl[c(2,4,6,8,10)], x1=xpos+0.5, y1=MuCrIr[c(2,4,6,8,10)],
lwd=1.2, col="darkorange")
segments(x0=xpos+0.35, y0=MuCrIl[c(2,4,6,8,10)], x1=xpos+0.65, y1=MuCrIl[c(2,4,6,8,10)],
lwd=1.2, col="darkorange")
segments(x0=xpos+0.35, y0=MuCrIr[c(2,4,6,8,10)], x1=xpos+0.65, y1=MuCrIr[c(2,4,6,8,10)],
lwd=1.2, col="darkorange")
mtext(expression(paste("Estimate of ", theta)), side = 3, cex = 1.2, font = 2, line = 0.5)
# Second plot:
plot(NA,ylab= expression(paste('Between-study variability (', tau, ")")),xlab='Model',
ylim= c(round(min(TauCrIl),2)-0.1, round(max(TauCrIr),2))+0.05,
xlim= c(-1, 13), xaxt='n',
family="serif", cex.axis=1.4, cex.lab=1.4,xaxs='i', yaxs='i')
#par(xpd=TRUE)
axis(side = 1, at = xpos, tick = T, labels = c("All (RC)", "All (RS)",
"Music (RC)", "Noise (RC)", "Speech (RC)"), padj = 0.5, cex=1.6)
#mtext("Model", line=4, side = 1, cex = 1.6, family="serif")
# Uniform means:
points(x= xpos-0.5, y= tauMean[c(1,3,5,7,9)], pch = 19, cex=1.2, col="darkred")
# Normal means:
points(x= xpos+0.5, y= tauMean[c(2,4,6,8,10)], pch = 15, cex=1.2, col="darkorange")
# Uniform CrIs:
segments(x0=xpos-0.5, y0=TauCrIl[c(1,3,5,7,9)], x1=xpos-0.5, y1=TauCrIr[c(1,3,5,7,9)],
lwd=1.2, col="darkred")
segments(x0=xpos-0.35, y0=TauCrIl[c(1,3,5,7,9)], x1=xpos-0.65, y1=TauCrIl[c(1,3,5,7,9)],
lwd=1.2, col="darkred")
segments(x0=xpos-0.35, y0=TauCrIr[c(1,3,5,7,9)], x1=xpos-0.65, y1=TauCrIr[c(1,3,5,7,9)],
lwd=1.2, col="darkred")
# Normal CrIs:
segments(x0=xpos+0.5, y0=TauCrIl[c(2,4,6,8,10)], x1=xpos+0.5, y1=TauCrIr[c(2,4,6,8,10)],
lwd=1.2, col="darkorange")
segments(x0=xpos+0.35, y0=TauCrIl[c(2,4,6,8,10)], x1=xpos+0.65, y1=TauCrIl[c(2,4,6,8,10)],
lwd=1.2, col="darkorange")
segments(x0=xpos+0.35, y0=TauCrIr[c(2,4,6,8,10)], x1=xpos+0.65, y1=TauCrIr[c(2,4,6,8,10)],
lwd=1.2, col="darkorange")
mtext(expression(paste("Estimate of ", tau)), side = 3, cex = 1.2, font = 2, line = 0.5)
dev.off()
#=========================================
# Leave-one-out method (robustness check):
#=========================================
source("Functions/sensitivity_analysis.R")
load("Data/subset/gen.Rda")
gen2<- gen[, c("g", "g_var")]
colnames(gen2)<- Bcols
# takes a lot of time
Sens1<- sensitivity_analysis(gen2, "dunif(-10, 10)", "dunif(0, 10)")
save(Sens1, file= "Summary/Sensitivity/G.Rda")
# reading speed
load("Data/subset/gen_speed.Rda")
gen_speed2<- gen_speed[, c('g', 'g_var')]
colnames(gen_speed2)<- Bcols
Sens2<- sensitivity_analysis(gen_speed2, "dunif(-10, 10)", "dunif(0, 10)")
save(Sens2, file= "Summary/Sensitivity/GR.Rda")
# proofreading accuracy:
load("Data/subset/proof.Rda")
proof2<- proof[, c('g', 'g_var')]
colnames(proof2)<- Bcols
Sens6<- sensitivity_analysis(proof2, "dunif(-10, 10)", "dunif(0, 10)")
save(Sens6, file= "Summary/Sensitivity/Proof.Rda")
# noise:
load("Data/subset/noise.Rda")
noise2<- noise[,c("g", "g_var")]
colnames(noise2)<- Bcols
Sens3<- sensitivity_analysis(noise2, "dunif(-10, 10)", "dunif(0, 10)")
save(Sens3, file= "Summary/Sensitivity/N.Rda")
# music:
load("Data/subset/music.Rda")
music2<- music[,c("g", "g_var")]
colnames(music2)<- Bcols
Sens4<- sensitivity_analysis(music2, "dunif(-10, 10)", "dunif(0, 10)")
save(Sens4, file= "Summary/Sensitivity/M.Rda")
# speech:
load("Data/subset/speech.Rda")
speech2<- speech[,c("g", "g_var")]
colnames(speech2)<- Bcols
Sens5<- sensitivity_analysis(speech2, "dunif(-10, 10)", "dunif(0, 10)")
save(Sens5, file= "Summary/Sensitivity/S.Rda")
#########################
# Alt. coding (LM) #
#########################
library(vioplot)
load("Posterior/LM1p.Rda")
load("Posterior/LM2p.Rda")
png('Plots/LM_alt_cod.png', width = 1600, height = 1600, units = "px", res = 300)
plot(1, 1, xlim = c(0.5,2.5), ylim = c(-1,1), type = 'n', xlab = 'Model',
ylab = 'Mean difference (g)', xaxt = 'n', family="serif", cex.lab=1.5, cex.axis=1.5,
main= "Lyrical vs non-lyrical music: \nMeta-regression results")
abline(a = 0,0, col="darkred", lwd=1.3, lty=2)
vioplot(LM1p, col=pallete[2], at=1, add=T)
vioplot(LM2p, col=pallete[3], at=2, add=T)
axis(1, at = c(1,2), labels = c('Paper', 'Alternative coding'), family="serif", cex.axis=1.3)
text(x = 1, y = 0.65, labels = paste("b= ", round(mean(LM1p),2), sep=""),
family="serif", cex= 1.2)
text(x = 2, y = 0.65, labels = paste("b= ", round(mean(LM2p),2), sep=""),
family="serif", cex= 1.2)
dev.off()