Odtnhj

I have been having trouble graphing growth curve model results that have been calculated over both subjects and items (both included as random effects), while models calculated over a dataset that is averaged over items, so that subjects are the only random effect, seems to work fine. I cannot seem to figure out why this would be or how to fix it.

Graph of Subject Only Model

Graph of Subjects and Items Model

Summary DF Subj Items

Summary DF Subj Only

> dput(head(new.df.subjitems, 20))
structure(list(Item.No = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 1), 
Subject_No = c(330, 322, 141, 330, 330, 
330, 137, 330, 330, 330, 137, 330, 137, 330, 330, 137, 137, 330, 
137, 141), 
Bin.No = c(35, 17, 19, 44, 42, 34, 31, 23, 36, 32, 
33, 28, 23, 33, 37, 7, 4, 30, 28, 31), 
TargetFix = c(1, 1, 1, 
0, 0.02, 1, 0, 1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0), 
Condition.E = structure(c(5L, 
5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
5L, 5L, 5L), .Label = c("First", "Second", "Max Entropy, Unrelated", 
"Third", "Fourth", "Max Entropy, Competitive"), class = "factor"), 


 ot1 = c(-0.00489995957550024, 0.220498180897511, -0.171498585142509, 
 0.142098827689507, 0.210698261746511, -0.20089834259551, 
 -0.0538995553305028, -0.151898746840508, -0.112699070236506, 
 -0.230298100048512, 0.0146998787265008, -0.161698665991508, 
 -0.151898746840508, 0.0146998787265008, -0.122498989387506, 
 0.0538995553305028, 0.181298504293509, -0.0440996361795023, 
 -0.161698665991508, -0.0538995553305028), 
ot2 = c(-0.158018948215706, 
 0.226392531578272, 0.0744512352170155, 0.00151941296361255, 
 0.192965446378795, 0.161057774142931, -0.135227753761518, 
 0.024310607417801, -0.0577376926172773, 0.26133902974136, 
 -0.156499535252094, 0.0486212148356019, 0.024310607417801, 
 -0.156499535252094, -0.0395047370539266, -0.135227753761518, 
 0.101800668562042, -0.142824818579581, 0.0486212148356019, 
 -0.135227753761518), 
ot3 = c(0.0112384799617412, 0.177762495548696, 
 0.0718614343707494, -0.14311051566666, 0.113069194230467, 
 -0.05612035827049, 0.113717568074414, 0.12506411034348, 0.167352493276442, 
 -0.250560469916257, -0.0335353360396831, 0.101434485808536, 
 0.12506411034348, -0.0335353360396831, 0.16389449944206, 
 -0.113717568074414, -0.035984748339037, 0.0957432042894495, 
 0.101434485808536, 0.113717568074414), 
ot4 = c(0.158779933129858, 
 0.0903598052498265, -0.175158477576023, -0.166766788637329, 
 -0.00086703192354873, -0.0706803112875428, 0.0864589810947309, 
 -0.178854426512889, -0.0950686572656207, 0.205524262921174, 
 0.153690832709029, -0.182263550144233, -0.178854426512889, 
 0.153690832709029, -0.123325047363878, 0.0864589810947308, 
 -0.15558879673071, 0.109609880754701, -0.182263550144233, 
 0.0864589810947309)), 
.Names = c("Item.No", "Subject_No", 
"Bin.No", "TargetFix", "Condition.E", "ot1", "ot2", "ot3", "ot4"
), row.names = c(1L, 5L, 8L, 22L, 29L, 59L, 61L, 74L, 78L, 86L, 
90L, 98L, 101L, 111L, 115L, 120L, 126L, 133L, 140L, 145L), class = "data.frame")


> dput(head(df.subjonly, 20))
structure(list(Subject_No = c(103, 103, 103, 103, 103, 103, 103, 
103, 103, 103, 103, 103, 103, 103, 103, 103, 103, 103, 103, 103
), 
Bin.No = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 
15, 16, 17, 18, 19, 20), 
TargetFix = c(0.16667, 0.16667, 0.16667, 
0.16667, 0.32667, 0.39, 0.5, 0.5, 0.5, 0.5, 0.62667, 0.59, 0.66667, 
0.66667, 0.76667, 0.76333, 0.66667, 0.40667, 0.33333, 0.48333
), 
Condition.E = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("First", 
"Second", "Max Entropy, Unrelated", "NaN, Unrelated", "Third", 
"Fourth", "Max Entropy, Competitive", "NaN, Competitive", "Low Entropy, NaN", 
"High Entropy, NaN", "Max Entropy, NaN", "NaN, NaN"), class = "factor"), 

ot1 = c(-0.240098019199512, -0.230298100048512, -0.220498180897511, 
 -0.210698261746511, -0.20089834259551, -0.19109842344451, 
 -0.181298504293509, -0.171498585142509, -0.161698665991508, 
 -0.151898746840508, -0.142098827689507, -0.132298908538507, 
 -0.122498989387506, -0.112699070236506, -0.102899151085505, 
 -0.0930992319345048, -0.0832993127835043, -0.0734993936325038, 
 -0.0636994744815033, -0.0538995553305028), 
ot2 = c(0.297804940868062, 
 0.26133902974136, 0.226392531578271, 0.192965446378795, 0.161057774142931, 
 0.13066951487068, 0.101800668562042, 0.0744512352170154, 
 0.0486212148356019, 0.0243106074178009, 0.0015194129636125, 
 -0.0197523685269634, -0.0395047370539266, -0.0577376926172774, 
 -0.0744512352170155, -0.0896453648531411, -0.103320081525654, 
 -0.115475385234555, -0.126111275979843, -0.135227753761518
 ), 
ot3 = c(-0.331823325024233, -0.250560469916256, -0.177762495548696, 
 -0.113069194230468, -0.0561203582704902, -0.00655577997768254, 
 0.0359847483390369, 0.0718614343707493, 0.101434485808536, 
 0.12506411034348, 0.14311051566666, 0.15593390946916, 0.16389449944206, 
 0.167352493276442, 0.166668098663387, 0.162201523293977, 
 0.154312974859294, 0.143362661050417, 0.12971078955843, 0.113717568074414
 ), 
ot4 = c(0.347265133901292, 0.205524262921176, 0.0903598052498273, 
 -0.000867031923547791, -0.0706803112875423, -0.121489365408538, 
 -0.15558879673071, -0.175158477576023, -0.182263550144233, 
 -0.178854426512889, -0.16676678863733, -0.147721588350685, 
 -0.123325047363878, -0.0950686572656211, -0.0643291795224186, 
 -0.0323686454785664, -0.000334356356151326, 0.0307411167449481, 
 0.0599399328470624, 0.0864589810947308)), .Names = c("Subject_No", 
"Bin.No", "TargetFix", "Condition.E", "ot1", "ot2", "ot3", "ot4"
), row.names = 3:22, class = "data.frame")
> 

# create 4th-order polynormial in the range of Bin.no
t <- poly ((unique(new.df.subjitems$Bin.No)), 4)
# create variables ot1, ot2, ot3, ot4 corresponding to the orthogonal 
polynomial time terms and populate their values
# with the Bin.No-appropriate orthogonal polynomial values:
new.df.subjitems[,paste("ot", 1:4, sep="")] <- t[new.df.subjitems$Bin.No, 1:4]
Model.subjitems.2 <- lmer(TargetFix ~ (ot1+ot2+ot3+ot4)*Condition.E + 
(1+ot1+ot2+ot3+ot4|Subject_No) +(1+ot1+ot2+ot3+ot4|Item.No),
 control = lmerControl(optimizer="bobyqa"), 
data=new.df.subjitems, REML=F)

# create 4th-order polynormial in the range of Bin.no
t <- poly ((unique(df.subjonly$Bin.No)), 4)
# create variables ot1, ot2, ot3, ot4 corresponding to the orthogonal 
polynomial time terms and populate their values
# with the Bin.No-appropriate orthogonal polynomial values:
df.subjonly[,paste("ot", 1:4, sep="")] <- t[df.subjonly$Bin.No, 1:4]
Model.subj.2 <- lmer(TargetFix ~ (ot1+ot2+ot3+ot4)*Condition.E + 
(1+ot1+ot2+ot3+ot4|Subject_No),
 control = lmerControl(optimizer="bobyqa"), data=df.subjonly, 
REML=F)

# Graph Subject Items
ggplot(data=new.df.subjitems, aes(Bin.No, TargetFix, color=Condition.E, 
lty=Condition.E, shape=Condition.E)) +
 stat_summary(fun.data=mean_se,geom="ribbon",linetype=0,alpha=0.25) +
 stat_summary(fun.data=mean_se,geom="point",size=1,alpha=0.40) 
+stat_summary(aes(y=fitted(Model.subjitems.2)),
fun.y=mean,geom="line",size=2.0,alpha=0.9) +
 theme_bw(base_size=10) +
 labs(y="Fixation Proportion", x="Bins") +
 ggtitle("Subjects and Items, Target Fixations") +
 theme(text=element_text(color = "black", size=20, family = "Georgia")) +
 theme(axis.text = element_text(color = "black", size=10, family = "Georgia")) 
+
 scale_color_viridis(begin=0, end = .8, discrete=TRUE) +
 scale_shape_manual(values=c(3,2,16,7)) + #16,3,2
 theme(legend.key.width=unit(4,"line")) +
 ggsave("Testing_SubjectItems_v3.png",width=10,height=5)

# Graph Subject Only
ggplot(data=df.subjonly, aes(Bin.No, TargetFix, color=Condition.E, 
lty=Condition.E, shape=Condition.E)) +
 stat_summary(fun.data=mean_se,geom="ribbon",linetype=0,alpha=0.25) +
 stat_summary(fun.data=mean_se,geom="point",size=1,alpha=0.40) +
stat_summary(aes(y=fitted(Model.subj.2)),
fun.y=mean,geom="line",size=2.0,alpha=0.9) +
 theme_bw(base_size=10) +
 labs(y="Fixation Proportion", x="Bins") +
 ggtitle("Subj Only, Target Fixations") +
 theme(text=element_text(color = "black", size=20, family = "Georgia")) +
 theme(axis.text = element_text(color = "black", size=10, family = "Georgia")) 
+
 scale_color_viridis(begin=0, end = .8, discrete=TRUE) +
 scale_shape_manual(values=c(3,2,16,7)) + #16,3,2
 theme(legend.key.width=unit(4,"line")) +
 ggsave("Testing_SubjectOnly_v3.png",width=10,height=5)