-
Notifications
You must be signed in to change notification settings - Fork 4
/
BMI trajectories VI V1
158 lines (129 loc) · 7.95 KB
/
BMI trajectories VI V1
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
#FIXED EFFECTS SPECIFICATION- SELECTING EFFECTS TO INCLUDE IN MIXED EFFECTS MODEL VIA GLM- BMI RESPONSE
library(lme4)
library(arm)
#ChildID causes problems in the model so has been removed but will be added to the VI model.
GLM1 <- glm(alldata$BMI ~ alldata$ageattest*alldata$Gender*alldata$Ethnicity+alldata$PostcodeDecile+alldata$SchoolCode+alldata$IsAccountRegistered+alldata$AcademicYear, data = alldata)
GLM2 <- glm(alldata$BMI ~ alldata$ageattest*alldata$Gender+alldata$Ethnicity+alldata$PostcodeDecile+alldata$SchoolCode+alldata$IsAccountRegistered+AcademicYear, data = alldata)
anova(GLM1, GLM2, test = "Chisq")
#INTERACTION SIGNIFICANT
GLM3 <- glm(alldata$BMI ~ alldata$ageattest+alldata$Gender*alldata$Ethnicity+alldata$PostcodeDecile+alldata$SchoolCode+alldata$IsAccountRegistered+AcademicYear, data = alldata)
anova(GLM1, GLM3, test = "Chisq")
#THREE WAY INTERACTION BETWEEN AGE, GENDER AND ETHNICITY
GLM4 <- glm(alldata$BMI ~ alldata$ageattest*alldata$Gender*alldata$Ethnicity+alldata$PostcodeDecile+alldata$SchoolCode+alldata$AcademicYear, data = alldata)
anova(GLM1, GLM4, test = "Chisq")
#CHAMP REGISTRATION NOT SIGNIFICANT
GLM5 <- glm(alldata$BMI ~ alldata$ageattest*alldata$Gender*alldata$Ethnicity+alldata$PostcodeDecile+alldata$SchoolCode, data = alldata)
anova(GLM1, GLM5, test = "Chisq")
#ACADEMIC YEAR IS SIGNIFICANT
GLM6 <- glm(alldata$BMI ~ alldata$ageattest*alldata$Gender*alldata$Ethnicity+alldata$PostcodeDecile+alldata$AcademicYear, data = alldata)
anova(GLM4, GLM6, test = "Chisq")
#SCHOOL CODE IS SIGNIFICANT
#all variables required for VI model
GLM4 <- glm(alldata$SDS_BMI ~ alldata$ageattest*alldata$Gender*alldata$Ethnicity+alldata$PostcodeDecile+alldata$SchoolCode+alldata$IsAccountRegistered, data = alldata)
GLc <- update(GLM1, subset = !is.na(alldata$AcademicYear))
anova(GLc, GLM4, test = "Chisq")
#Academic year is not significant, so doesn't need to be included
GLM5 <- glm(alldata$SDS_BMI ~ alldata$ageattest*alldata$Gender*alldata$Ethnicity+alldata$PostcodeDecile+alldata$SchoolCode, data = alldata)
GLc2 <- update(GLM4, subset = !is.na(alldata$IsAccountRegistered))
anova(GLc2, GLM5, test = "Chisq")
#Academic year is not significant, so doesn't need to be included
#Options for anova tests= Rao (Rao's efficient score), LRT (likelihood ratio test), F, Chisq, Cp (Mallow's Cp statistic)
#Postcode decile is important to include.
summary(GLM2)
summary(GLM2b)
#Postcode produces a simpler (see df) model than school code, explaining more of the deviance.
#Postcode should be included as a fixed effect, and school possibly as a random effect
GLM2 <- glm(alldata$SDS_BMI ~ alldata$ageattest + alldata$Gender + alldata$Ethnicity + alldata$PostcodeDecile, data = alldata)
GLM3 <- glm(alldata$SDS_BMI ~ alldata$ageattest + alldata$Gender + alldata$PostcodeDecile, data = alldata)
GLc<- update(GLM3, subset = !is.na(alldata$Ethnicity))
anova(GLc, GLM2, test="Chisq")
#Ethnicity is important and needs to be in the final model (include interaction with social decile)
GLM4<- glm(alldata$SDS_BMI ~ alldata$ageattest + alldata$Gender + alldata$Ethnicity + alldata$PostcodeDecile +alldata$AcademicYear, data = alldata)
GLc<- update(GLM2, subset = !is.na(alldata$AcademicYear))
anova(GLc, GLM4, test = "Chisq")
#CHAMP is significant
#VARIABLE INTERCEPTS MODEL
VI1<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + subsad$Ethnicity*subsad$PostcodeDecile+subsad$IsAccountRegistered+(1|subsad$ChildID)+(1|subsad$SchoolCode),REML=FALSE)
summary(VI1)
#model converges
VI1b<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + subsad$Ethnicity+subsad$PostcodeDecile+subsad$IsAccountRegistered+(1|subsad$ChildID)+(1|subsad$SchoolCode),REML=FALSE)
anova(VI1,VI1b,test="LRT")
#The interaction between ethnicity and postcode decile is important
VI2<- lmer(subsad$SDS_BMI~ subsad$ageattest+subsad$Gender + subsad$Ethnicity*subsad$PostcodeDecile+subsad$IsAccountRegistered+(1|subsad$ChildID)+(1|subsad$SchoolCode),REML=FALSE)
anova(VI1,VI2,test="LRT")
#The interaction between age and gender is important
VI3<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + subsad$Ethnicity*subsad$PostcodeDecile+(1|subsad$ChildID)+(1|subsad$SchoolCode),REML=FALSE)
anova(VI1,VI3,test="LRT")
#IsAccountRegistered is not a significant predictor in this dataset
#probably because it depends on the point of registration
VI4<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + subsad$Ethnicity*subsad$PostcodeDecile+(1|subsad$ChildID),REML=TRUE)
anova(VI3,VI4,test="LRT")
#School is important as a random effect
VI5<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + subsad$Ethnicity*subsad$PostcodeDecile+(1|subsad$SchoolCode),REML=FALSE)
anova(VI3,VI5,test="LRT")
#ChildID is a significant predictor as a random effect
#VI3 gives the best fit. ChildID improves fit (based on AIC), as do the interactions between age and gender, and ethnicity and postcode decile.
Ethnicity2<-subsad$Ethnicity
levels(Ethnicity2)
levels(Ethnicity2)[1:2]<-"Other"
levels(Ethnicity2)[17:18]<-"Other"
VI6<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity2*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI3, VI6,test="LRT")
#count unknown ethnicity as part of other
Ethnicity3<-Ethnicity2
levels(Ethnicity3)[5:6]<-"Mixed black/white"
VI7<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity3*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI6, VI7,test="LRT")
#mixed black african and mixed black caribbean can be grouped
Ethnicity4<-Ethnicity3
levels(Ethnicity4)[7]<-"Other"
VI8<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity4*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI7, VI8,test="LRT")
#'any other mixed background' can be grouped with other
Ethnicity5<-Ethnicity4
levels(Ethnicity5)
levels(Ethnicity5)[2:4]<-"White"
VI9<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity5*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI8, VI9,test="LRT")
Ethnicity5<-Ethnicity4
levels(Ethnicity5)[2:3]<-"White"
VI9<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity5*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI8, VI9,TEST="LRT")
#British and Irish can be grouped but not with other white
Ethnicity6<-Ethnicity5
levels(Ethnicity6)
levels(Ethnicity6)[3]<-"White other"
levels(Ethnicity6)[6]<-"Indian/Bangladeshi"
levels(Ethnicity6)[8]<-"Indian/Bangladeshi"
VI10<-lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity6*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI9, VI10,test="LRT")
#Indian doesn't group with Bangladeshi
Ethnicity8<-Ethnicity5
levels(Ethnicity8)[9]<-"Other Asian"
levels(Ethnicity8)[7]<-"Other Asian"
VI11<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity8*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI9, VI11,test="LRT")
#group mixed asian with Pakistani
Ethnicity9<-Ethnicity8
levels(Ethnicity9)[10]<-"Black"
levels(Ethnicity9)[11]<-"Black"
VI12<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity9*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI11, VI12,test="LRT")
#Black African and other black can be grouped
Ethnicity10<-Ethnicity9
levels(Ethnicity10)[9]<-"Black"
VI13<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity10*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI12, VI13,test="LRT")
#Carib background can be grouped with other black but only just
Ethnicity11<-Ethnicity10
levels(Ethnicity11)[10]<-"Chinese"
levels(Ethnicity11)[3]<-"Other white"
levels(Ethnicity11)[6]<-"Indian"
levels(Ethnicity11)[8]<-"Bangladeshi"
levels(Ethnicity11)[5]<-"Other Asian"
VI14<- lmer(subsad$SDS_BMI~ subsad$ageattest*subsad$Gender + Ethnicity11*subsad$PostcodeDecile +(1|subsad$ChildID)+(1|subsad$School),REML=FALSE)
anova(VI13, VI14,test="LRT")
#mixed Asian cant be grouped with other Asian
subsad$Ethnicity10<-Ethnicity10
summary(VI13)
print(VI13,correlation=TRUE)