Data Harmonization: Establishing Measurement Invariance across Different Assessments of the Same Construct across Adolescence
- 11 June 2019
- journal article
- research article
- Published by Taylor & Francis in Journal of Clinical Child & Adolescent Psychology
- Vol. 48 (4) , 555-567
Abstract
Longitudinal measurement invariance is a major concern for developmental scholars who seek to evaluate the same underlying construct across time. Unfortunately, discontinuities in the expression of various psychological constructs, as well as essential changes in measurement that are necessitated by shifting developmental capacities and practice effects over time, make the task of establishing longitudinal invariance extremely difficult. Drawing on 5 waves of longitudinal data from 392 families (52% female; Mage_W1 = 12.89, SD = .48; Mage_W5 = 21.95, SD = .77; 199 European American and 193 Mexican American families), the current investigation sought to establish measurement invariance across developmentally appropriate changes in measures of depressive symptomatology from early adolescence through early adulthood. Using a combination of item parceling and the common and unique items from 2 assessment instruments for depressive symptoms, the data supported strong factorial invariance in youth’s depressive symptoms across 5 waves of measurement. Findings suggest that traditional invariance approaches can be adapted to determine whether the same construct underlies different measurement instruments across time. This analytic strategy can allow researchers and clinicians to use more sophisticated techniques to understand changes in symptomatology regardless of changes in measurement or developmental capacity. Applying this approach to model patterns of depressive symptomatology from early adolescence to early adulthood has important clinical implications for elucidating periods when youth experience elevations in depressive symptoms and heightened needs for intervention services.Keywords
Funding Information
- Ford Foundation (Not Applicable)
- National Institute of Child Health and Human Development (R01HD0566-06A1)
- National Institute of Mental Health (MH64829 R01; T32 MH015755)
This publication has 36 references indexed in Scilit:
- Why item parcels are (almost) never appropriate: Two wrongs do not make a right—Camouflaging misspecification with item parcels in CFA models.Psychological Methods, 2013
- Why the items versus parcels controversy needn’t be one.Psychological Methods, 2013
- Factorial Invariance Within Longitudinal Structural Equation Models: Measuring the Same Construct Across TimeChild Development Perspectives, 2010
- Relations between Mattering to Step- and Non-Residential Fathers and Adolescent Mental HealthFathering: A Journal of Theory, Research, and Practice about Men as Fathers, 2009
- Modeling life-span growth curves of cognition using longitudinal data with multiple samples and changing scales of measurement.Psychological Methods, 2009
- An Empirical Evaluation of the Use of Fixed Cutoff Points in RMSEA Test Statistic in Structural Equation ModelsSociological Methods & Research, 2008
- Pooling data from multiple longitudinal studies: The role of item response theory in integrative data analysis.Developmental Psychology, 2008
- To Parcel or Not to Parcel: Exploring the Question, Weighing the MeritsStructural Equation Modeling: A Multidisciplinary Journal, 2002
- Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternativesStructural Equation Modeling: A Multidisciplinary Journal, 1999
- IRT Equating MethodsEducational Measurement: Issues and Practice, 1991