Simulation Study of Linking Using Mirt

R
mirt
linking
Author

Jihong Zhang

Published

November 10, 2017

This simulation study is to show how to do IRT Linking Process using mirt R Package. The simulation data includes 2 forms - Form A and Form B. These 2 forms are simulated based on 2 groups of individual, one group has 0 mean trait, another has 0.25 mean trait. Both groups have same sd.

2 Introduction

The means and sds of simulated Form A and B are like:

\theta_{A} = \theta_{B} - 0.25 \\ \sigma_A^2 = \sigma_B^2

3 Calibration of Form A

The mean of \theta for individuals administrated with form A is 0, the standard deviation (SD=1). In the dataset, X is ID, V1 is true trait (\theta), V3 to V52 is unique items, V54 to V63 are common items.

3.1 Look at the data

First of all, have a look at the data

R Code 
library(mirt)
library(tidyverse)
Warning: package 'dplyr' was built under R version 4.2.3
Warning: package 'stringr' was built under R version 4.2.3
R Code 
library(knitr)
### Read in Raw data from Form A:
dat <- read.csv(file="FormA.csv")
glimpse(dat)
Rows: 5,000
Columns: 64
$ X   <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19,…
$ V1  <dbl> -0.79498, -0.05589, 0.62650, -1.26832, 0.74921, 1.00922, -0.19147,…
$ V2  <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA…
$ V3  <int> 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, …
$ V4  <int> 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, …
$ V5  <int> 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 0, 1, …
$ V6  <int> 0, 0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, …
$ V7  <int> 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, …
$ V8  <int> 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 1, 0, …
$ V9  <int> 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1, …
$ V10 <int> 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, …
$ V11 <int> 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 0, 1, 1, 0, …
$ V12 <int> 0, 1, 1, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, …
$ V13 <int> 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, …
$ V14 <int> 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0, …
$ V15 <int> 0, 0, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, …
$ V16 <int> 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, …
$ V17 <int> 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
$ V18 <int> 0, 0, 1, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, 0, 0, …
$ V19 <int> 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, …
$ V20 <int> 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, …
$ V21 <int> 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ V22 <int> 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, …
$ V23 <int> 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, …
$ V24 <int> 0, 1, 1, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 1, …
$ V25 <int> 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, …
$ V26 <int> 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, …
$ V27 <int> 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, …
$ V28 <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 0, 1, …
$ V29 <int> 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, …
$ V30 <int> 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1, 1, 0, …
$ V31 <int> 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 0, …
$ V32 <int> 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, …
$ V33 <int> 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1, 0, …
$ V34 <int> 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, …
$ V35 <int> 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 0, …
$ V36 <int> 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, …
$ V37 <int> 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, …
$ V38 <int> 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, …
$ V39 <int> 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, …
$ V40 <int> 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, …
$ V41 <int> 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, …
$ V42 <int> 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, …
$ V43 <int> 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, …
$ V44 <int> 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, …
$ V45 <int> 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 1, 1, 1, …
$ V46 <int> 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 1, 1, 0, …
$ V47 <int> 0, 0, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, …
$ V48 <int> 0, 0, 1, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, …
$ V49 <int> 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 0, 1, 1, 0, 0, …
$ V50 <int> 0, 1, 0, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, …
$ V51 <int> 0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, …
$ V52 <int> 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, …
$ V53 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA…
$ V54 <int> 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 0, …
$ V55 <int> 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, …
$ V56 <int> 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, …
$ V57 <int> 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, …
$ V58 <int> 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 0, 0, 1, 1, 0, …
$ V59 <int> 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, …
$ V60 <int> 0, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, …
$ V61 <int> 1, 0, 0, 0, 1, 1, 0, 1, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0, …
$ V62 <int> 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, …
$ V63 <int> 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, …

3.2 Plot the density of true \theta of Group A

From the density function, mu_{\theta} is 0, sd_{\theta} is 1.

R Code 
plot(density(dat$V1), main = "Group A True Trait Density", 
     xlab=expression(theta) )

3.3 CTT Table

CTT table could provide a brief description of table. 2 key valables in the table is item difficulty (item.diff) calculated by item means P(y=1) and item discrimination (item.disc), which is item-total correlation.

3.3.1 Clean data

By cleaning, data has 60 items including 50 unique item (from item1 to item50) and 10 common items (from item51 to item60). The sample size is 5000.

R Code 
dat_cali <- dat %>% select(V3:V52, V54:V63)
colnames(dat_cali) <- paste0("item",1:60)
N <- nrow(dat_cali)
n <- ncol(dat_cali)

3.3.2 Classical Test Theory

Then calculate the CTT table for Form A. The item discrimination and difficulty could be compared between Form A and Form B. Because the relationship between trait and total score is non-linear, so there is effect of shrinkage.

R Code 
# item stats
## item discrimnation
item.disc <- apply(dat_cali, 2, function(x) cor(x, rowSums(dat_cali, na.rm = TRUE)))

## item difficulty
item.diff <- colMeans(dat_cali)

## item response frequency
item.freq <- reduce(lapply(dat_cali, table), bind_rows)

CTT <- cbind(item.disc, item.diff, item.freq)

kable(CTT, digits = 3, caption = "CTT Table for Form A")
CTT Table for Form A
item.disc item.diff 0 1
item1 0.365 0.761 1197 3803
item2 0.464 0.833 835 4165
item3 0.457 0.539 2305 2695
item4 0.471 0.524 2380 2620
item5 0.385 0.221 3893 1107
item6 0.400 0.563 2183 2817
item7 0.261 0.457 2717 2283
item8 0.264 0.365 3176 1824
item9 0.331 0.560 2198 2802
item10 0.505 0.431 2843 2157
item11 0.306 0.354 3228 1772
item12 0.528 0.560 2199 2801
item13 0.362 0.550 2251 2749
item14 0.427 0.761 1196 3804
item15 0.433 0.241 3793 1207
item16 0.377 0.458 2710 2290
item17 0.524 0.545 2275 2725
item18 0.351 0.578 2112 2888
item19 0.429 0.509 2455 2545
item20 0.330 0.307 3465 1535
item21 0.399 0.403 2984 2016
item22 0.418 0.609 1957 3043
item23 0.356 0.354 3232 1768
item24 0.460 0.670 1651 3349
item25 0.388 0.507 2465 2535
item26 0.209 0.310 3451 1549
item27 0.242 0.476 2618 2382
item28 0.495 0.502 2492 2508
item29 0.441 0.481 2593 2407
item30 0.450 0.731 1344 3656
item31 0.601 0.399 3006 1994
item32 0.396 0.682 1588 3412
item33 0.584 0.613 1933 3067
item34 0.489 0.557 2216 2784
item35 0.467 0.759 1203 3797
item36 0.312 0.367 3166 1834
item37 0.410 0.778 1109 3891
item38 0.443 0.571 2143 2857
item39 0.519 0.755 1224 3776
item40 0.280 0.327 3365 1635
item41 0.413 0.582 2088 2912
item42 0.338 0.395 3024 1976
item43 0.430 0.634 1830 3170
item44 0.348 0.382 3088 1912
item45 0.370 0.530 2351 2649
item46 0.286 0.337 3314 1686
item47 0.456 0.403 2985 2015
item48 0.340 0.397 3014 1986
item49 0.310 0.291 3544 1456
item50 0.435 0.357 3214 1786
item51 0.283 0.821 895 4105
item52 0.272 0.324 3382 1618
item53 0.474 0.531 2347 2653
item54 0.298 0.336 3318 1682
item55 0.427 0.546 2271 2729
item56 0.471 0.667 1667 3333
item57 0.410 0.470 2652 2348
item58 0.483 0.433 2834 2166
item59 0.363 0.232 3841 1159
item60 0.376 0.666 1670 3330

3.4 Final Calibration of Form A

3.4.1 Model Specification

R Code 
SPECS <- mirt.model('F = 1-60
                    PRIOR = (1-60, a1, lnorm, 0, 1),
                    (1-60,  d,  norm, 0, 1),
                    (1-60,  g,  norm, -1.39,1)')
mod_A3PL <- mirt(data=dat_cali, model=SPECS, itemtype='3PL')
parms_a <- coef(mod_A3PL, simplify=TRUE, IRTpars = TRUE)$items
R Code 
a_A <- parms_a[,1] 
b_A <- parms_a[,2] 
c_A <- parms_a[,3] 
theta_A <- fscores(mod_A3PL,method="EAP",
                   full.scores=TRUE, full.scores.SE=TRUE,
                   scores.only=TRUE)

Using 3-PL for irt model of form A. Extracting the cofficients (a, b, c) of IRT. The model-implied theta was outputed, whose mean is 0.1314938

The plot below suggest that SE is low when theta is close to mean, but low theta and high theta has large SE.

R Code 
plot(theta_A[,1],theta_A[,2])

4 Calibration of Form B

The structure of Form B is as same as Form A.

R Code 
dat_b <- read.csv(file="FormB.csv")
glimpse(dat_b)
Rows: 5,000
Columns: 64
$ X   <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19,…
$ V1  <dbl> 1.12513, 0.29724, -0.54719, 1.18477, 0.37517, 0.05821, 1.11918, 1.…
$ V2  <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA…
$ V3  <int> 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 1, 0, …
$ V4  <int> 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, …
$ V5  <int> 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, …
$ V6  <int> 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, …
$ V7  <int> 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 1, 1, 1, 0, …
$ V8  <int> 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, …
$ V9  <int> 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 1, 1, 1, 1, 1, …
$ V10 <int> 1, 0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 1, 0, …
$ V11 <int> 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 1, …
$ V12 <int> 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, …
$ V13 <int> 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, …
$ V14 <int> 0, 1, 0, 1, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, …
$ V15 <int> 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, …
$ V16 <int> 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 0, …
$ V17 <int> 0, 0, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, …
$ V18 <int> 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, …
$ V19 <int> 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, …
$ V20 <int> 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, …
$ V21 <int> 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, …
$ V22 <int> 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, …
$ V23 <int> 1, 1, 0, 1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, …
$ V24 <int> 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, …
$ V25 <int> 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 1, 0, …
$ V26 <int> 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 1, …
$ V27 <int> 1, 0, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, …
$ V28 <int> 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0, 0, 1, 1, 1, 0, …
$ V29 <int> 0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 1, …
$ V30 <int> 1, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, …
$ V31 <int> 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, …
$ V32 <int> 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, …
$ V33 <int> 0, 0, 0, 0, 1, 0, 0, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, …
$ V34 <int> 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, …
$ V35 <int> 1, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, …
$ V36 <int> 0, 1, 0, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 0, …
$ V37 <int> 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, …
$ V38 <int> 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1, …
$ V39 <int> 1, 1, 0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, …
$ V40 <int> 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, …
$ V41 <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 0, …
$ V42 <int> 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1, …
$ V43 <int> 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, …
$ V44 <int> 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, …
$ V45 <int> 1, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ V46 <int> 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, …
$ V47 <int> 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, …
$ V48 <int> 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, …
$ V49 <int> 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, …
$ V50 <int> 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, …
$ V51 <int> 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, …
$ V52 <int> 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, …
$ V53 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA…
$ V54 <int> 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, …
$ V55 <int> 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, …
$ V56 <int> 0, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, …
$ V57 <int> 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, …
$ V58 <int> 1, 1, 0, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, …
$ V59 <int> 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, …
$ V60 <int> 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, …
$ V61 <int> 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, …
$ V62 <int> 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, …
$ V63 <int> 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, …
R Code 
library(ggplot2)

ggplot(data = dat_b) +
  ggtitle("Group B True Trait Density") +
  geom_density(aes(x = V1), fill = "skyblue", col = "skyblue", alpha=0.8) +
  xlab("True Latent Trait")

R Code 
dat_cali2 <- dat_b %>% select(V3:V52, V54:V63)
colnames(dat_cali2) <- paste0("item",1:60)
N <- nrow(dat_cali2)
n <- ncol(dat_cali2)


# item stats
## item discrimnation
item.disc <- apply(dat_cali2, 2, function(x) cor(x, rowSums(dat_cali, na.rm = TRUE)))

## item difficulty
item.diff <- colMeans(dat_cali2)

## item response frequency
item.freq <- reduce(lapply(dat_cali2, table), bind_rows)

CTT2 <- cbind(item.disc, item.diff, item.freq)

kable(CTT2, digits = 3, caption = "CTT Table for Form A")
CTT Table for Form A
item.disc item.diff 0 1
item1 -0.029 0.517 2415 2585
item2 -0.005 0.573 2137 2863
item3 -0.037 0.552 2242 2758
item4 0.008 0.519 2405 2595
item5 0.013 0.709 1455 3545
item6 -0.012 0.804 979 4021
item7 0.017 0.687 1567 3433
item8 0.003 0.641 1794 3206
item9 0.000 0.626 1869 3131
item10 -0.021 0.392 3038 1962
item11 0.021 0.821 893 4107
item12 -0.012 0.474 2629 2371
item13 -0.026 0.687 1564 3436
item14 0.008 0.517 2414 2586
item15 0.004 0.475 2626 2374
item16 -0.012 0.525 2375 2625
item17 -0.025 0.363 3185 1815
item18 0.001 0.393 3036 1964
item19 0.003 0.490 2550 2450
item20 0.001 0.505 2477 2523
item21 0.006 0.710 1451 3549
item22 0.015 0.248 3759 1241
item23 0.007 0.594 2030 2970
item24 -0.010 0.360 3200 1800
item25 0.011 0.664 1681 3319
item26 -0.002 0.790 1050 3950
item27 0.011 0.395 3026 1974
item28 0.009 0.353 3236 1764
item29 -0.003 0.335 3325 1675
item30 -0.015 0.346 3268 1732
item31 0.005 0.340 3301 1699
item32 0.007 0.613 1935 3065
item33 -0.026 0.377 3117 1883
item34 0.003 0.500 2502 2498
item35 -0.007 0.436 2821 2179
item36 -0.009 0.767 1166 3834
item37 0.003 0.428 2861 2139
item38 0.002 0.593 2033 2967
item39 0.011 0.663 1683 3317
item40 0.020 0.606 1968 3032
item41 -0.010 0.840 799 4201
item42 0.015 0.554 2232 2768
item43 0.012 0.824 880 4120
item44 -0.018 0.433 2837 2163
item45 -0.005 0.429 2855 2145
item46 -0.019 0.316 3421 1579
item47 -0.007 0.827 867 4133
item48 -0.023 0.760 1201 3799
item49 0.009 0.389 3055 1945
item50 -0.013 0.317 3413 1587
item51 -0.002 0.848 762 4238
item52 -0.018 0.362 3191 1809
item53 0.007 0.580 2101 2899
item54 -0.013 0.363 3184 1816
item55 -0.023 0.581 2096 2904
item56 -0.034 0.731 1344 3656
item57 -0.007 0.524 2379 2621
item58 -0.017 0.491 2547 2453
item59 0.004 0.275 3627 1373
item60 0.022 0.715 1423 3577

4.1 Final Calibration of Form A

4.1.1 Model Specification of B

R Code 
SPECS2 <- mirt.model('F = 1-60
                    PRIOR = (1-60, a1, lnorm, 0, 1),
                    (1-60,  d,  norm, 0, 1),
                    (1-60,  g,  norm, -1.39,1)')
mod_B3PL <- mirt(data=dat_cali2, model=SPECS, itemtype='3PL')
parms_b <- coef(mod_B3PL, simplify=TRUE, IRTpars = TRUE)$items
R Code 
a_B <- parms_b[,1] 
b_B <- parms_b[,2] 
c_B <- parms_b[,3] 
theta_B <- fscores(mod_B3PL,method="EAP",
                   full.scores=TRUE, full.scores.SE=TRUE,
                   scores.only=TRUE)
head(theta_B, 20) %>% kable(digits = 3, caption = "Model-implied Theta of B")
Model-implied Theta of B
F SE_F
0.933 0.214
0.193 0.220
-0.783 0.293
0.865 0.208
-0.233 0.261
-0.217 0.239
0.890 0.215
1.369 0.232
-0.875 0.306
-0.411 0.266
1.091 0.223
-0.638 0.261
-1.253 0.321
-0.784 0.261
0.436 0.228
0.135 0.211
-1.841 0.491
-0.491 0.259
0.065 0.225
-0.339 0.240

4.2 b-plot

The relationship between b parameters of A and B reflect the latent traits of A and B:

\theta_{A} = \theta_{B} - 0.25 \\ \sigma_A^2 = \sigma_B^2 thus, b_B -b_A should also be -0.25. the estimated difference of b is calculate by the mean of b parametes of Form A’s common items and that of Form B’s common items, which is -0.2756443. Thus, it is very close to difference of true traits.

R Code 
### b-plot
plot(b_A[51:60],b_B[51:60],
     main=paste0("r =", round(cor(b_A[51:60],b_B[51:60]),5)), 
     xlab = "b_A",
     ylab = "b_B"
     )

R Code 
# mean(b_B[51:60])-mean(b_A[51:60])

4.3 a-plot

Because \theta_B - \theta_A = 0.25, so a parametes are: a_A / a_B = 1

The true ratio of (means of) a parameters is 1.016284, which is very close to 1. SD of A and B are both close to 1.

R Code 
### a-plot
plot(a_A[51:60],a_B[51:60],
     main=round(cor(a_A[51:60],a_B[51:60]),5)
     )

R Code 
# mean(a_A[51:60]) / mean(a_B[51:60]) 


#SDs of b-values across forms
SD_bA <- sd(b_A[51:60])
SD_bB <- sd(b_B[51:60])

Mean_bA <- mean(b_A[51:60])
Mean_bB <- mean(b_B[51:60])

5 Linking

R Code 
###
### Run one or the other, NOT BOTH!
###

### MS linking: place item parameters from
### Form B on the scale of Form A
slope <- SD_bA / SD_bB
inter <- Mean_bA - slope*Mean_bB

### MM linking: place item parameters from
### Form B on the scale of Form A
# slope <- Mean_aA / Mean_aB
# inter <- Mean_bA - slope*Mean_bB


###
### Perform the Linking
###

LINKED_items <- matrix(0,50,3)

#Column 3 is c, it stays the same
LINKED_items[,3] <- c_B[1:50]

#Column 2 is b, it is linked:
LINKED_items[,2] <- b_B[1:50]*slope + inter

#Column 1 is a, it is also linked:
LINKED_items[,1] <- a_B[1:50] / slope

### Now the ITEM BANK has 110 items all
### linked to a common metric!
rownames(LINKED_items) <- paste0("item_B", 1:50)
ITEM.BANK <- rbind(parms_a[,-4], LINKED_items)
# write.csv(ITEM.BANK,file="item_bank.csv")

#Link the theta estimates
LINKED_theta <- theta_B[,1]*slope + inter
LINKED_se <- theta_B[,2]/slope

# write.csv(cbind(LINKED_theta,LINKED_se),file="LINKED_FormB_theta_est.csv")
paste0("Mean of Theta of B is ",mean(theta_B[,1]) %>% round(3))
[1] "Mean of Theta of B is -0.013"
R Code 
paste0("SD of Theta of B is ", sd(theta_B[,1]) %>% round(3))
[1] "SD of Theta of B is 0.97"
R Code 
# after Linked
print("After Linking:")
[1] "After Linking:"
R Code 
paste0("Mean of Theta of B is ",mean(LINKED_theta) %>% round(3))
[1] "Mean of Theta of B is 0.269"
R Code 
paste0("SD of Theta of B is ",sd(LINKED_theta) %>% round(3))
[1] "SD of Theta of B is 0.935"
R Code 
# Theta of A is
paste0("Mean of Theta of A is ",mean(theta_A[,1]) %>% round(3))
[1] "Mean of Theta of A is -0.021"
R Code 
paste0("SD of Theta of A is ", sd(theta_A[,1]) %>% round(3))
[1] "SD of Theta of A is 0.965"
Back to top