Transcript Differential Item Functioning by Ethnicity in the Inhibitory Control

Differential Item Functioning by Ethnicity in the Inhibitory Control Subscale of the
Children’s Behavior Questionnaire (CBQ)
Kristin L. Moilanen, Chad B. Henry, Daniel S. Shaw, Thomas J. Dishion, Frances Gardner, &
Melvin N. Wilson
Methods
•720 families enrolled in the Early Steps Multisite Project, a multi-site prevention
program for young children at increased risk for clinically-elevated levels of
conduct problems.
•Child demographics:
•Sex: 50% male
•Race/ethnicity: 46% White, 26% Black, 28% Hispanic/Other
•Site: 25% Charlottesville, VA, 38% Eugene, OR, & 37% Pittsburgh, PA
•Primary caregivers (97% mothers) completed the inhibitory control subscale of
the CBQ about their child at the age 4 assessment.
•Inhibitory control : 7-point response scale, ranging from 1 (extremely untrue of
child) to 7 (extremely true of child). Mothers could also indicate whether any items
were not applicable to their child. Scale scores were computed by averaging all
numeric responses.
Results & Discussion
•Group differences in scale total and item means are presented in Table 1.
•There were no ethnic differences in overall inhibitory control, but there were
mean level differences by ethnic group for 5 of the 13 scale items.
•Effect sizes were not constant across items, which is a rough indicator that DIF is
present (Smith, 2002).
Black
Hispanic/Other
N
291
166
171
Cronbach’s α
.79
.64
.74
Scale M (SD)
4.43 (.82)
4.36 (.77)
4.55 (.82)
Cohen’s d
W/B
Cohen’s d
W/HO
Cohen’s d
B/HO
.09
-.15
-.24
Item M(SD)
1
Can lower voice
5.18 (1.48) O
5.38 (1.58)
5.53 (1.35) W
-.13
-.25
-.10
2
Is good at games
5.22 (1.40) O
5.40 (1.50)
5.65 (1.36) W
-.12
-.31
-.17
3
Hard time following instructions 4.10 (1.54)
3.77 (1.66)
4.12 (1.68)
.21
-.01
-.21
4
Prepares for trips
4.54 (1.70)
4.63 (1.62)
4.86 (1.57)
-.05
-.20
-.14
5
Wait before starting activities
4.82 (1.42) O
4.73 (1.64) O
5.11 (1.41) W, B
.06
-.20
-.25
6
Difficulty waiting in line
3.71 (1.61)
3.62 (1.77)
3.75 (1.73)
.05
-.02
-.07
7
Trouble sitting still
3.43 (1.69)
3.19 (1.69)
3.40 (1.77)
.14
.02
-.12
8
Able to resist laughing
3.90 (1.53)
3.97 (1.78)
3.86 (1.81)
-.04
.02
.06
9
Good at following instructions
5.17 (1.26) B
4.83 (1.62) W, O
5.19 (1.33) B
.23
-.01
-.24
10 Approaches slowly
4.61 (1.91) B
4.16 (1.99) W, O
4.74 (1.93) B
.23
-.06
-.29
11 Not very careful
4.32 (1.93)
4.49 (2.02)
4.22 (2.01)
-.08
.05
.13
12 Can easily stop
4.35 (1.50)
4.39 (1.65)
4.52 (1.68)
-.03
-.10
-.08
13 Resist temptation
4.39 (1.47)
4.25 (1.63)
4.29 (1.65)
.09
.06
-.02
Note. Bold values indicate significant mean differences.
DIF Analyses
DIF was detected by comparing nested ordinal regression models (Zumbo, 1999).
Seven items did not show DIF by ethnicity. These items are relatively unambiguous
and correspond to the “core elements” of inhibitory control during preschool
years.
•4. Prepares for trips and outings by planning things s/he will need.
•5. Can wait before entering new activities if s/he is asked to do so.
•6. Has difficulty waiting in line for something. (REV)
•7. Has trouble sitting still when s/he is told to (at movies, church, etc.). (REV)
•8. Is able to resist laughing or smiling when it isn’t appropriate.
•11. Is not very careful and cautious in crossing streets. (REV)
•12. Can easily stop an activity when s/he is told “no.”
Item 7 - Does not show DIF
Item 10 - Shows DIF
0.9
0.7
0.5
0.3
0.1
-0.1
0.9
W1
W7
B1
B7
H1
H7
-3
-2
-1
0
1
2
3
Standard Deviations Above/Below the Mean
Total Inhibitory Control Score
Response Probability
Study Purpose
The purpose of the current study was to explore ethnicity as a source of
differential item functioning (DIF) in the inhibitory control subscale of the
Children’s Behavior Questionnaire (CBQ; Rothbart, Ahadi, Hershey, & Fisher, 2001).
This questionnaire was initially developed using samples of European American
children, and has since been empirically validated in many nations around the
world. Yet, it is problematic that potential item-level differences between ethnic
groups have not yet been fully examined, because different ethnic groups may
define children’s self-regulation in unique ways which may not be accurately
represented in the current scale (i.e., presumed ethnic differences could actually
be measurement bias).
White
Response Probability
Abstract
In this study, the 13-item inhibitory control subscale of the Children’s Behavior
Questionnaire (CBQ) was examined for evidence of differential item functioning
(DIF) by ethnicity. As part of a larger study, 720 primary caregivers completed this
subscale when their children were 4 years old. Nested ordinal regression analyses
indicated that 7 items functioned equivalently across ethnic groups. Four of these
items favored White children, and two favored children of Black and
Hispanic/Other ethnicities. Further attention to this subscale’s item properties is
necessary in order to advance research on the development of inhibitory control
in diverse populations of children.
Table 1
Descriptive Statistics
0.7
0.5
0.3
0.1
-0.1
W1
W4
W7
B1
B4
B7
-3
-2
-1
0
1
2
3
Standard Deviations Above/Below the Mean
Total Inhibitory Control Score
Acknowledgements
This research was supported by grant 5 R01 DA16110 from the National Institutes of Health to the third and fourth authors.
We gratefully thank the Early Steps and the Child and Family Center staff for their assistance, and all the families who
participated in this project.
Six items demonstrated DIF by ethnicity. Four of these items favored White
children, and two favored both Black and Hispanic/Other children.
Three of the four items that favored White children over Black children concerned
compliance with instructions.
•3. Has a hard time following instructions. (REV)
•9. Is good at following instructions.
•10. Approaches places s/he has been told are dangerous slowly and
cautiously.
Mothers of Black children may have higher expectations for their preschoolers’
compliance than the mothers of White children.
White children were also favored over Hispanic/Other children on item 13 (“Is
usually able to resist temptation when told s/he is not supposed to do
something”). This is the only item that includes the modifier “usually.” White and
Hispanic/Other mothers may have different standards for what it means for their
children to “usually” resist temptation.
Black and Hispanic/Other children were favored over EA children on two items
that did not share a common theme.
•1. Can lower his/her voice when asked to do so.
•2. Is good at games like “Simon Says,” “Mother, May I?” and “Red Light,
Green Light.”
Parents of White children may have higher expectations for children’s speaking
volume. Although item 2 makes theoretical sense, it is possible that mothers
respond about their child’s overall game-playing skills, not their ability to play
these particular games thought to involve inhibitory control.
Black children were also favored over Hispanic/Other children on item 10 (see text
above). The features and accessibility of “dangerous places” may vary for urban
African American mothers and suburban/rural Hispanic/Other mothers.
Conclusions & Future Directions
•Nearly half of the items in this subscale do not function equivalently across ethnic
groups.
•No one group was consistently favored over the others by the items that showed
DIF.
•DIF may be reduced by sharpening item wording and using more concrete scale
anchors.
•As at least three items reference compliance to instructions, it might be beneficial
to eliminate at least one of these items so compliance isn’t over-represented in
the scale.
•More work is needed to determine if parents of different ethnicities have
divergent expectations for children’s “controlled” behavior and emotions, and
different perceptions of what “control” looks like.