The Disaggregation of Within-Person and Between-Person Effects in Longitudinal Models of Change

Abstract

Longitudinal models are becoming increasingly prevalent in the behavioral sciences, with key advantages including increased power, more comprehensive measurement, and establishment of temporal precedence. One particularly salient strength offered by longitudinal data is the ability to disaggregate between-person and within-person effects in the regression of an outcome on a timevarying covariate. However, the ability to disaggregate these effects has not been fully capitalized upon in many social science research applications. Two likely reasons for this omission are the general lack of discussion of disaggregating effects in the substantive literature and the need to overcome several remaining analytic challenges that limit existing quantitative methods used to isolate these effects in practice. This review explores both substantive and quantitative issues related to the disaggregation of effects over time, with a particular emphasis placed on the multilevel model. Existing analytic methods are reviewed, a general approach to the problem is proposed, and both the existing and proposed methods are demonstrated using several artificial data sets. Potential limitations and directions for future research are discussed, and recommendations for the disaggregation of effects in practice are offered.

Development and validation of a 6-item working alliance questionnaire for repeated administrations during psychotherapy.

“…Curran & Bauer, 2011). This means that some response-set biases that are stable over time but vary among patients, such as social desirability responding and acquiescence (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) can be separated from fluctuations in alliance quality from one session to another within the same therapy.…”

“…The two-step procedure was used in order to isolate the within-patient effects and to control for non-stationarity, issues that may bias results if not accounted for (Curran & Bauer, 2011). In the present study, we instead used the Autoregressive Latent Trajectory model (Bollen & Curran, 2004;Curran & Bollen, 2001) Working Alliance, autoregression and residuals were estimated separately for each occasion.…”

Perceived Racial Discrimination and Negative-Mood–Related Drinking Among African American College Students

“…Daily mood was person-mean centered so that effects represented differences in drinking level based on deviations from each individual's average mood over the study month (Enders & Tofighi, 2007). Furthermore, the person-mean for daily negative mood was included as a predictor to allow us to separate within-person and between-person effects of daily mood on drinking (Curran & Bauer, 2011). Finally, we included a Daily Negative Mood × Discrimination × Gender interaction term (and all relevant two-way interactions) to test whether individuals who reported higher levels of discrimination showed stronger associations between daily negative mood and that evening's level of drinking, and whether such relations differed between men and women.…”

The Temporal Order of Changes in Physical Activity and Subjective Sleep in Depressed Versus Nondepressed Individuals: Findings From the MOOVD Study

“…Therefore, trends were removed from the variables using ordinary least squares (OLS) regression, for each individual separately, before the analyses were performed. In order to distinguish effects within participants from effects between participants, the recommended approach of "personmean centering of the predictor variables" was used by calculating the daily deviation from the participants' mean values for the predictors (Bolger & Laurenceau, 2013;Curran & Bauer, 2011). As a result the mean value of each predictor equaled zero for all participants, canceling out all between-158 subjects differences in levels.…”