CODEWITHSUNDEEP
rsimulation
Tabea
Tabea

Reputation: 67

Generate data with simsem package in R

I would like to generate data with the simsem package in R. So far, I've done this:

    # create data with simsem package 
    ## factor loadings    
    loading <- matrix(0, 7)
    loading[1,1] <- NA
    LY2 <- bind(loading, 1)
     
    ## residual correlation matrix
    latent.cor <- matrix(NA, 1, 1)
    diag(latent.cor) <- 1
    RPS2 <- binds(latent.cor, 0.5)
    
    # measurement error corr matrix
    RTE2 <- binds(diag(7))

    # total variance of indicators    
    VY2 <- bind(rep(NA,7),2)
    
    CFA.Model2 <- model(LY = LY2, RPS = RPS2, RTE = RTE2, modelType = "CFA")
    
    # generate data 
    dist <- bindDist("norm", list(mean = 4, sd = 1))
    dat <- generate(CFA.Model2, n = 50, indDist = dist)

I was hoping to create item responses for 7 items that range from 1 to 5. The data that I create however looks like this:

head(dat, 4)
          y1           y2           y3         y4          y5
1 -0.3042082 -0.009703124  0.070651822  1.9138537 -0.02754102
2  0.9574723 -1.691375825  0.441645186 -0.1770509 -0.35793280
3 -2.1808565  1.467395026 -0.350395973 -0.1660219 -0.42191898
4 -0.2367881  0.003594693 -0.002771362 -0.1323401 -1.44860960
          y6          y7
1 -1.2557268 -1.52874138
2 -0.9963241  0.87807237
3 -1.7527848  0.31383091
4  1.2102580  0.03469505

How can I create data that ranges from 1 to 5? I would want the result to be the following:

head(dat, 4)
   y1  y2  y3  y4  y5  y6  y7
1   3   5   3   5   5   5   5
2   5   5   5   5   5   3   5
3   3   4   1   2   2   2   4
4   5   4   1   2   4   5   4

Upvotes: 2

Views: 169

Answers (2)

Tabea
Tabea

Reputation: 67

I solved it by specifying the indicator means:

### create data with simsem package 

## factor loadings   
loading <- matrix(NA, 7)
loading2 <- matrix("runif(1, 0, 0.5)", 7)
LY2 <- bind(loading, loading2)

## residual correlation matrix
latent.cor <- matrix(NA, 1, 1)
diag(latent.cor) <- "1"
RPS2 <- binds(latent.cor)

## measurement error corr matrix
RTE2 <- binds(diag(7))

## total variance of indicators
var <- matrix("runif(1, 0, 2)", 7)
VY2 <- bind(rep(NA,7), var)

## indicator means
indmeans <- matrix("runif(1, 2, 3)", 7)
MY <- bind(rep(NA,7), indmeans)

CFA.model <- model(MY = MY, LY = LY2, RPS = RPS2, RTE = RTE2, modelType = "CFA")

# Draw a parameter set for data generation.
param <- draw(CFA.model)

# create data
dat <- createData(param[[1]], n = 50) 

b1 <- round(dat[,1], 0)
b2 <- round(dat[,2], 0)
b3 <- round(dat[,3], 0)
b4 <- round(dat[,4], 0)
b5 <- round(dat[,5], 0)
b6 <- round(dat[,6], 0)
b7 <- round(dat[,7], 0)

Upvotes: 1

Ronak Shah
Ronak Shah

Reputation: 389095

You can do this in base R using replicate function to generate data randomly between 1 to 5.

n <- 10
replicate(7, sample(5, n, replace = TRUE))

#      [,1] [,2] [,3] [,4] [,5] [,6] [,7]
# [1,]    5    3    3    3    1    4    3
# [2,]    5    3    4    5    2    5    1
# [3,]    1    3    1    2    1    5    1
# [4,]    4    5    3    1    5    3    3
# [5,]    4    4    1    3    2    3    2
# [6,]    3    1    2    4    5    5    4
# [7,]    1    4    1    2    5    4    4
# [8,]    5    5    3    5    4    4    1
# [9,]    2    1    3    1    1    5    4
#[10,]    1    5    4    2    5    1    5

Upvotes: 0

Related Questions