Manipulation of Intensive Longitudinal Data: A Tutorial in R With Applications on the Job Demand‐Control Model

Intensive longitudinal designs (ILD) are increasingly used in applied psychology to investigate research questions and deliver interventions at both within- and between-individual levels. However, while relatively complex analyses such as cross-level interaction models are trending in the field, little guidance has been provided on ILD data manipulation, including all procedures to be applied to the raw data points for getting the final dataset to be analysed. Here, we provide an introductory step-by-step tutorial and open-source R code on required and recommended data pre-processing (e.g., data reading, merging and cleaning), psychometric (e.g., level-specific reliability), and other ILD data manipulation procedures (e.g., data centering, lagging and leading). We built our tutorial on an illustrative example aimed at testing the job demand-control model at the within-individual level based on data from 211 back-office workers who received up to 18 surveys over three workdays, supporti...

Intensive longitudinal designs (ILD) are increasingly used in applied psychology to investigate research questions and deliver interventions at both within- and between-individual levels. However, while relatively complex analyses such as cross-level inter-action models are trending in the field, little guidance has been provided on ILD data manipulation, including all procedures to be applied to the raw data points for getting the final dataset to be analysed. Here, we provide an introductory step-by-step tutorial and open-source R code on required and recommended data pre-processing (e.g., data reading, merging and cleaning), psycho-metric (e.g., level-specific reliability), and other ILD data manipulation procedures (e.g., data centering, lagging and leading). We built our tutorial on an illustrative example aimed at testing the job demand-control model at the within-individual level based on data from 211 back-office workers who received up to 18 surveys over three workdays, supporting both the strain and (partially) the buffer hypotheses. Being the common starting point of many types of analyses, data manipulation is crucial to determine the quality and validity of the resulting study outcomes. Hence, this tutorial and the attached code aim to contribute to removing methodological barriers among applied psychology researchers and practitioners in the handling of ILD data.