Identifying Unbiased Items for Screening Preschoolers for Disruptive Behavior Problems.

Area of Science:

• Child Psychology
• Pediatric Behavioral Health
• Psychometrics

Background:

• Early identification of disruptive behavior problems is vital for timely intervention.
• Existing screening tools are often unsuitable for diverse pediatric primary care populations.
• Measurement bias in screening instruments can lead to inequitable identification of behavioral issues.

Purpose of the Study:

• To identify measurement bias (differential item functioning, DIF) in candidate items for a new preschool behavioral screening tool.
• To assess DIF by child race, sex, and socioeconomic status.
• To refine a brief screening instrument for diverse pediatric populations.

Main Methods:

• Examined 18 candidate items for a new brief behavioral screening instrument.
• Utilized item response theory (IRT) likelihood ratio tests and ordinal logistic regression to detect DIF.
• Analyzed data from 900 parents/guardians of preschool-aged children across four primary care settings.

Main Results:

• Eight of the 18 candidate items exhibited statistically significant differential item functioning (DIF).
• Bias was identified across child race, sex, and socioeconomic status for these eight items.
• The presence of DIF indicated these items were not functioning equivalently across different child subgroups.

Conclusions:

• The identified bias in 8 of 18 items renders them inappropriate for screening diverse child populations.
• These biased items were excluded from the development of the new brief screening tool.
• This study contributes to the development of more equitable and valid behavioral screening instruments for primary care.