First, let’s get some wordbank data for American English. We’ll start with production data from the CDI: Words & Gestures (WG) form. We’ll look for DIF based on 1) sex and 2) SES (high vs. low). Questions: How do we choose a dozen anchor items (that we expect to be unbiased)?

Let’s do comprehension data.

None Primary Some Secondary Secondary Some College College Some Graduate Graduate
0 2 68 254 266 292 41 140
Var1 Freq
action_words 55
animals 36
body_parts 20
clothing 19
descriptive_words 37
food_drink 30
furniture_rooms 24
games_routines 19
household 36
locations 11
outside 27
people 20
pronouns 11
quantifiers 8
question_words 6
sounds 12
time_words 8
toys 8
vehicles 9

We will fit comprehension data from 2385 children for the WG form.

Fit baseline 2PL model

GLIMMER Plots

## Warning: ggrepel: 58 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps

DIF Model: Sex

## Joining, by = c("a1", "definition")
## Joining, by = c("a1", "definition")

Items with DIF in difficulty

275 items show DIF based on a \(\chi^2\) test yielding \(p<.05\). Below we show the 171 items with \(p<.001\).

AIC BIC X2 p
doll -326.34 -320.57 328.34 0
dress (object) -283.44 -277.67 285.44 0
pretty -133.86 -128.08 135.86 0
truck -113.42 -107.64 115.42 0
hammer -96.28 -90.51 98.28 0
necklace -84.62 -78.85 86.62 0
purse -79.91 -74.13 81.91 0
same -75.60 -69.82 77.60 0
soap -74.63 -68.85 76.63 0
baby -70.64 -64.86 72.64 0
wash -69.95 -64.17 71.95 0
hair -69.27 -63.50 71.27 0
leg -68.75 -62.97 70.75 0
face -67.19 -61.41 69.19 0
sleep -66.94 -61.16 68.94 0
shoe -66.59 -60.82 68.59 0
see -66.16 -60.38 68.16 0
naughty -65.89 -60.11 67.89 0
read -65.62 -59.85 67.62 0
cute -65.58 -59.80 67.58 0
asleep -65.38 -59.60 67.38 0
chair -65.22 -59.45 67.22 0
help -65.15 -59.38 67.15 0
show -64.98 -59.20 66.98 0
flower -64.82 -59.04 66.82 0
put -64.74 -58.96 66.74 0
picture -64.48 -58.70 66.48 0
hi -64.28 -58.51 66.28 0
clean (action) -63.98 -58.20 65.98 0
dance -63.97 -58.20 65.97 0
tummy -63.84 -58.06 65.84 0
tickle -63.77 -58.00 65.77 0
blanket -63.52 -57.75 65.52 0
teddybear -63.52 -57.74 65.52 0
sweater -63.38 -57.60 65.38 0
little (description) -63.35 -57.57 65.35 0
girl -63.24 -57.47 65.24 0
glasses -63.19 -57.41 65.19 0
clean (description) -63.12 -57.34 65.12 0
school -63.11 -57.34 65.11 0
tongue -63.07 -57.29 65.07 0
me -63.07 -57.29 65.07 0
eye -62.99 -57.21 64.99 0
mouse -62.91 -57.13 64.91 0
pen -62.90 -57.12 64.90 0
bed -62.77 -56.99 64.77 0
beads -62.67 -56.90 64.67 0
cheek -62.66 -56.89 64.66 0
when -62.54 -56.77 64.54 0
play pen -62.52 -56.75 64.52 0
walk -62.37 -56.60 64.37 0
you -62.29 -56.51 64.29 0
dirty -62.27 -56.50 64.27 0
paper -62.04 -56.26 64.04 0
coat -62.01 -56.23 64.01 0
potty -61.98 -56.20 63.98 0
bottle -61.81 -56.03 63.81 0
knee -61.76 -55.98 63.76 0
child -61.74 -55.96 63.74 0
cat -61.68 -55.90 63.68 0
happy -61.67 -55.89 63.67 0
who -61.63 -55.85 63.63 0
hello -61.62 -55.85 63.62 0
meow -61.59 -55.81 63.59 0
cry -61.58 -55.80 63.58 0
play -61.54 -55.76 63.54 0
fish (food) -61.52 -55.75 63.52 0
yes -61.52 -55.74 63.52 0
keys -61.48 -55.70 63.48 0
take -61.39 -55.61 63.39 0
finger -61.36 -55.58 63.36 0
zipper -61.35 -55.57 63.35 0
your -61.31 -55.54 63.31 0
tooth -61.31 -55.53 63.31 0
I -61.27 -55.49 63.27 0
finish -61.23 -55.45 63.23 0
cup -61.22 -55.45 63.22 0
night night -61.18 -55.40 63.18 0
bib -61.10 -55.32 63.10 0
swim -61.06 -55.28 63.06 0
babysitter’s name -61.05 -55.27 63.05 0
bye -60.78 -55.00 62.78 0
stroller -60.66 -54.88 62.66 0
pattycake -60.24 -54.46 62.24 0
bunny -60.22 -54.44 62.22 0
comb -60.15 -54.37 62.15 0
please -60.13 -54.35 62.13 0
tomorrow -60.11 -54.33 62.11 0
nice -60.08 -54.30 62.08 0
yum yum -60.05 -54.28 62.05 0
love -59.96 -54.18 61.96 0
red -59.96 -54.18 61.96 0
belly button -59.94 -54.17 61.94 0
sing -59.94 -54.17 61.94 0
rocking chair -59.94 -54.17 61.94 0
mine -59.91 -54.13 61.91 0
kiss -59.90 -54.12 61.90 0
good -59.89 -54.11 61.89 0
write -59.88 -54.10 61.88 0
money -59.84 -54.07 61.84 0
shh/shush/hush -59.80 -54.02 61.80 0
candy -59.80 -54.02 61.80 0
thank you -59.77 -54.00 61.77 0
lady -59.76 -53.99 61.76 0
later -59.74 -53.97 61.74 0
grandma* -59.74 -53.97 61.74 0
party -59.71 -53.93 61.71 0
firetruck -57.81 -52.03 59.81 0
broom -48.77 -43.00 50.77 0
shovel -32.80 -27.02 34.80 0
motorcycle -30.41 -24.64 32.41 0
train -27.11 -21.34 29.11 0
vroom -26.20 -20.42 28.20 0
vacuum -22.85 -17.07 24.85 0
bathroom -20.86 -15.09 22.86 0
touch -19.51 -13.73 21.51 0
choo choo -19.49 -13.71 21.49 0
kitchen -19.36 -13.58 21.36 0
towel -19.28 -13.50 21.28 0
mouth -16.56 -10.78 18.56 0
home -15.94 -10.16 17.94 0
pajamas -15.84 -10.06 17.84 0
get -15.51 -9.73 17.51 0
arm -15.44 -9.67 17.44 0
dry (description) -15.15 -9.38 17.15 0
out -14.96 -9.18 16.96 0
foot -14.83 -9.06 16.83 0
inside -14.67 -8.90 16.67 0
bowl -14.52 -8.75 16.52 0
brush -14.15 -8.37 16.15 0
open -14.12 -8.34 16.12 0
sun -13.73 -7.96 15.73 0
bread -13.68 -7.91 15.68 0
stop -13.67 -7.89 15.67 0
bear -13.63 -7.85 15.63 0
toe -13.24 -7.46 15.24 0
there -13.12 -7.35 15.12 0
store -13.05 -7.27 15.05 0
hurry -12.78 -7.00 14.78 0
wanna/want to -12.59 -6.82 14.59 0
big -12.57 -6.80 14.57 0
couch -12.54 -6.76 14.54 0
nap -12.22 -6.44 14.22 0
button -12.12 -6.34 14.12 0
pillow -11.91 -6.13 13.91 0
table -11.67 -5.89 13.67 0
airplane -11.46 -5.69 13.46 0
bird -11.14 -5.36 13.14 0
hungry -11.12 -5.35 13.12 0
toy (object) -10.96 -5.18 12.96 0
hat -10.76 -4.98 12.76 0
plate -10.75 -4.97 12.75 0
milk -10.61 -4.84 12.61 0
bring -10.55 -4.77 12.55 0
diaper -10.54 -4.77 12.54 0
shorts -10.48 -4.70 12.48 0
squirrel -10.23 -4.45 12.23 0
give -10.16 -4.38 12.16 0
all -9.86 -4.09 11.86 0
bus -9.84 -4.06 11.84 0
apple -9.83 -4.05 11.83 0
ride -9.62 -3.85 11.62 0
head -9.57 -3.80 11.57 0
bee -9.53 -3.75 11.53 0
pig -9.43 -3.65 11.43 0
animal -9.41 -3.63 11.41 0
quack quack -9.39 -3.61 11.39 0
turtle -9.37 -3.60 11.37 0
deer -9.31 -3.54 11.31 0
noodles -8.91 -3.14 10.91 0
fall -8.91 -3.13 10.91 0

Items with a large difference in difficulty across sex

definition a1 d_m d_f d_diff d_diff_abs
dress (object) 2.20 -4.58 -1.41 3.17 3.17
doll 1.75 -1.56 0.87 2.43 2.43
when 2.63 -7.63 -5.96 1.67 1.67
pretty 1.52 -2.29 -0.74 1.55 1.55
hammer 2.07 -2.22 -3.69 -1.47 1.47
necklace 2.00 -3.36 -1.90 1.46 1.46
purse 2.05 -2.17 -0.85 1.32 1.32
tomorrow 2.59 -7.15 -5.91 1.24 1.24
truck 2.09 0.16 -1.04 -1.20 1.20
old 2.84 -7.69 -6.58 1.11 1.11
sweater 2.05 -3.61 -2.53 1.08 1.08
baby 1.72 0.73 1.76 1.04 1.04
bye 2.05 3.97 4.99 1.03 1.03
firetruck 2.12 -1.84 -2.85 -1.01 1.01
girl 1.95 -2.87 -1.87 1.00 1.00
chair 3.21 -0.48 0.46 0.94 0.94
tonight 1.96 -5.49 -4.57 0.92 0.92
child 1.72 -4.07 -3.15 0.92 0.92
shovel 2.08 -2.75 -3.66 -0.91 0.91
beads 1.55 -3.59 -2.71 0.87 0.87
cute 1.49 -2.52 -1.65 0.87 0.87
same 2.85 -6.64 -5.78 0.86 0.86
motorcycle 1.89 -2.57 -3.41 -0.84 0.84
later 2.17 -4.23 -3.39 0.84 0.84
naughty 1.16 -3.29 -2.45 0.84 0.84
tooth 2.54 -0.91 -0.07 0.84 0.84
broom 2.12 -1.04 -1.87 -0.83 0.83
hi 1.43 1.66 2.48 0.82 0.82
belly button 2.07 -0.60 0.21 0.81 0.81
little (description) 2.96 -4.38 -3.58 0.80 0.80
leg 3.13 -2.52 -1.73 0.79 0.79
shoe 3.11 2.16 2.95 0.78 0.78
rocking chair 1.94 -2.19 -1.41 0.78 0.78
picture 2.74 -2.06 -1.28 0.78 0.78
bed 2.57 -0.03 0.75 0.78 0.78
paper 2.50 -1.97 -1.19 0.78 0.78
hair 3.10 0.34 1.12 0.77 0.77
help 2.82 -2.67 -1.90 0.77 0.77
kiss 1.83 1.71 2.48 0.76 0.76
lady 2.02 -3.79 -3.03 0.76 0.76
flower 2.67 -0.90 -0.14 0.76 0.76
eye 2.53 0.41 1.17 0.76 0.76
red 1.96 -4.45 -3.70 0.75 0.75
party 2.13 -4.91 -4.17 0.74 0.74
write 2.05 -3.56 -2.82 0.74 0.74
clean (description) 2.55 -2.59 -1.86 0.74 0.74
clean (action) 2.57 -2.01 -1.28 0.73 0.73
please 2.01 -1.32 -0.60 0.72 0.72
money 1.96 -3.07 -2.36 0.71 0.71

Plot item ease for males vs. females

We label the 49 items with absolute ease difference of more than median + 1SD = 0.71.

Anchor items

For our DIF analysis, let’s pick 10 anchor items that show the smallest difficulty difference in the slope-invariant free group means model (mg3).

definition a1 d_f d_m d_diff d_diff_abs
jeans 2.512 -4.489 -4.490 0.001 0.001
his 2.300 -4.305 -4.302 -0.003 0.003
hit 1.995 -0.979 -0.982 0.003 0.003
plant 2.126 -1.915 -1.909 -0.006 0.006
pants 2.942 -1.357 -1.350 -0.008 0.008
ball 2.812 3.516 3.507 0.009 0.009
push 2.018 -1.199 -1.213 0.014 0.014
light 2.324 -0.328 -0.311 -0.016 0.016
park 1.798 -1.698 -1.715 0.017 0.017
stairs 2.083 -0.636 -0.658 0.022 0.022

Ability distribution comparison

What are the ability distributions by sex in the 2PL model?

## Joining, by = "data_id"

## Joining, by = "data_id"

sex ability_2pl ability_sex
Female 0.22 0.05
Male 0.02 0.05

Average male ability is slightly higher in the sex group model (i.e., there is a larger female advantage in the standard 2PL model), showing that the 36 items that are easier for females boost female ability estimates.

Ability distribution comparison

What are the ability distributions by SES in the 2PL model?

## Joining, by = c("data_id", "ability")

## Joining, by = "data_id"

ses_group ability_2pl ability_ses
high -0.13 0.24
low -0.17 0.05