cdriveraus

##install ctsem (If after May 2019, CRAN is recommended)
install.packages("devtools")
library(devtools)
install_github("cdriveraus/ctsem")
library(ctsem)


#specify generative model (linear sde only at present for data generation -- on the to do list)
gm <- ctModel(
  type='omx', #omx is older model type still needed for data generation

cdriveraus

n=100
a<-rnorm(n)
b<-a*.5 + rnorm(n)
bs <- scale(b)
as<-scale(a)

summary(lm(b~a))
summary(lm(a~b))
summary(lm(bs~as))
summary(lm(as~bs))

cdriveraus

N=10000 #iterations
obs=20 #observations per iteration
beta=exp(rnorm(N,-2,1))

Bs=c() #bias in estimated Beta (standardised) for no error condition
Bsm=c() #bias in estimated Beta (standardised) for with error condition
betas2=c() #unstandardised true beta for no error iterations that came out significant
betasm2=c() #unstandardised true beta for with error iterations that came out significant

for(i in 1:N){

cdriveraus

n=10000
beta=.4
#no effect

p=c()
for(i in 1:n){
s=rnorm(30,0,1)
sm=s+rnorm(30,0,1) #with measurement error
x=rnorm(30,0,1)
p=c(p,summary(lm(sm~x))$coefficients[2,4])

cdriveraus

library("qgraph")
library("igraph")
library("IsingSampler")
library("IsingFit")
library('rstan')

set.seed(1337)
Kappa <- as.matrix(get.adjacency(watts.strogatz.game(1,10,1,0)))
Kappa[1,5] <- Kappa[5,1] <- Kappa[3,9] <- Kappa[9,3] <-Kappa[7,1] <-Kappa[1,7] <- Kappa[5,9] <-Kappa[9,5] <-Kappa[7,3] <-Kappa[3,7] <-.5

cdriveraus

b=.8
n=100
nruns=10000
out<-rep(NA,nruns)

for(i in 1:nruns){
x=rnorm(n,3,2)
y=b*x + rnorm(n, 0, 3)

fit=lm(y~x)