Investigating which type of maltreatment has the largest impact on health: The challenge of measurement error
Abstract
To answer the question of which type of maltreatment has the largest impact on health outcomes, researchers often simultaneously enter multiple maltreatment predictors in a single regression model. However, this approach of mutual adjustment is problematic, as some maltreatment types (e.g. physical neglect) are often measured with more error than others (e.g. physical abuse). Greater measurement error attenuates effect sizes and reduces statistical power to detect associations, implicitly creating the appearance that better measured maltreatment types are more impactful. The attenuation is particularly severe in mutually adjusted models, as maltreatment types tend to correlate with each other. Researchers comparing maltreatment types should routinely report on reliability estimates such as McDonald’s omega and present separately-estimated associations next to the mutually adjusted model.
Related
- The impact of contamination bias when replicating maltreatment effects on child behavior problems
- Investigating the Relationship Between Mitral Leaflet Separation Index and Severity of Mitral Valve Stenosis Using Two and Three-Dimensional Measurement Methods in Trans Esophageal Echocardiography.
- Investigating the Causal Impact of COVID-19 on Acute Myocardial Infarction: A Multi-ancestry Mendelian Randomization Study
- The impact of meditation on cognitive flexibility: A critical review and the case for investigating True Self Meditation
- Investigating the impact of allicin on periodontal tissue in rats with periodontitis associated with type 2 diabetes mellitus
BibTeX
@misc{zarchev2026investigating,
title = {{Investigating which type of maltreatment has the largest impact on health: The challenge of measurement error}},
author = {Milan Zarchev and Henning Tiemeier},
howpublished = {PsyArXiv},
year = {2026},
month = sep,
doi = {10.31234/osf.io/dh3r2_v2},
url = {https://osf.io/preprints/psyarxiv/dh3r2_v2/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Investigating which type of maltreatment has the largest impact on health: The challenge of measurement error
AU - Milan Zarchev
AU - Henning Tiemeier
PY - 2026
DA - 2026/09/24
PB - PsyArXiv
DO - 10.31234/osf.io/dh3r2_v2
UR - https://osf.io/preprints/psyarxiv/dh3r2_v2/
AB - To answer the question of which type of maltreatment has the largest impact on health outcomes, researchers often simultaneously enter multiple maltreatment predictors in a single regression model. However, this approach of mutual adjustment is problematic, as some maltreatment types (e.g. physical neglect) are often measured with more error than others (e.g. physical abuse). Greater measurement error attenuates effect sizes and reduces statistical power to detect associations, implicitly creating the appearance that better measured maltreatment types are more impactful. The attenuation is particularly severe in mutually adjusted models, as maltreatment types tend to correlate with each other. Researchers comparing maltreatment types should routinely report on reliability estimates such as McDonald’s omega and present separately-estimated associations next to the mutually adjusted model.
ER -