This article presents a strategy for the initial step of data harmonization in Individual Participant Data syntheses, i.e., making decisions as to which measures operationalize the constructs of interest - and which do not. This step is vital in the process of data harmonization, because a study can only be as good as its measures. If the construct validity of the measures is in question, study results are questionable as well. Our proposed strategy for data harmonization consists of three steps. First, a unitary construct is defined based on the existing literature, preferably on the theoretical framework surrounding the construct. Second, the various instruments used to measure the construct are evaluated as operationalizations of this construct, and retained or excluded based on this evaluation. Third, the scores of the included measures are recoded on the same metric. We illustrate the use of this method with three example constructs focal to the Collaboration on Attachment Transmission Synthesis (CATS) study: parental sensitivity, child temperament, and social support. This process description may aid researchers in their data pooling studies, filling a gap in the literature on the first step of data harmonization.Data harmonization in studies using combined datasets is of vital importance for the validity of the study results.We have developed and illustrated a strategy on how to define a unitary construct and evaluate whether instruments are operationalizations of this construct as the initial step in the harmonization process.This strategy is a transferable and reproducible method to apply to the data harmonization process. (C) 2022 The Author(s). Published by Elsevier B.V.
Conceptual comparison of constructs as first step in data harmonization: Parental sensitivity, child temperament, and social support as illustrations
Cassibba, R.;Coppola, G.;Costantini, A.;
2022-01-01
Abstract
This article presents a strategy for the initial step of data harmonization in Individual Participant Data syntheses, i.e., making decisions as to which measures operationalize the constructs of interest - and which do not. This step is vital in the process of data harmonization, because a study can only be as good as its measures. If the construct validity of the measures is in question, study results are questionable as well. Our proposed strategy for data harmonization consists of three steps. First, a unitary construct is defined based on the existing literature, preferably on the theoretical framework surrounding the construct. Second, the various instruments used to measure the construct are evaluated as operationalizations of this construct, and retained or excluded based on this evaluation. Third, the scores of the included measures are recoded on the same metric. We illustrate the use of this method with three example constructs focal to the Collaboration on Attachment Transmission Synthesis (CATS) study: parental sensitivity, child temperament, and social support. This process description may aid researchers in their data pooling studies, filling a gap in the literature on the first step of data harmonization.Data harmonization in studies using combined datasets is of vital importance for the validity of the study results.We have developed and illustrated a strategy on how to define a unitary construct and evaluate whether instruments are operationalizations of this construct as the initial step in the harmonization process.This strategy is a transferable and reproducible method to apply to the data harmonization process. (C) 2022 The Author(s). Published by Elsevier B.V.File | Dimensione | Formato | |
---|---|---|---|
PIIS2215016122002680-2.pdf
accesso aperto
Tipologia:
Documento in Versione Editoriale
Licenza:
Creative commons
Dimensione
782.27 kB
Formato
Adobe PDF
|
782.27 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.