A critique of the cross-lagged panel model.
TL;DR: This article presents an alternative model that separates the within-person process from stable between-person differences through the inclusion of random intercepts, and discusses how this model is related to existing structural equation models that include cross-lagged relationships.
Abstract: The cross-lagged panel model (CLPM) is believed by many to overcome the problems associated with the use of cross-lagged correlations as a way to study causal influences in longitudinal panel data. The current article, however, shows that if stability of constructs is to some extent of a trait-like, time-invariant nature, the autoregressive relationships of the CLPM fail to adequately account for this. As a result, the lagged parameters that are obtained with the CLPM do not represent the actual within-person relationships over time, and this may lead to erroneous conclusions regarding the presence, predominance, and sign of causal influences. In this article we present an alternative model that separates the within-person process from stable between-person differences through the inclusion of random intercepts, and we discuss how this model is related to existing structural equation models that include cross-lagged relationships. We derive the analytical relationship between the cross-lagged parameters from the CLPM and the alternative model, and use simulations to demonstrate the spurious results that may arise when using the CLPM to analyze data that include stable, trait-like individual differences. We also present a modeling strategy to avoid this pitfall and illustrate this using an empirical data set. The implications for both existing and future cross-lagged panel research are discussed.
Citations
More filters
TL;DR: It is shown that the cross-lagged estimates derived from the ARCL model reflect a weighted-and typically uninterpretable-amalgam of between- and within-person associations.
Abstract: Reciprocal feedback processes between experience and development are central to contemporary developmental theory. Autoregressive cross-lagged panel (ARCL) models represent a common analytic approach intended to test such dynamics. The authors demonstrate that—despite the ARCL model's intuitive appeal—it typically (a) fails to align with the theoretical processes that it is intended to test and (b) yields estimates that are difficult to interpret meaningfully. Specifically, using a Monte Carlo simulation and two empirical examples concerning the reciprocal relation between spanking and child aggression, it is shown that the cross-lagged estimates derived from the ARCL model reflect a weighted—and typically uninterpretable—amalgam of between- and within-person associations. The authors highlight one readily implemented respecification that better addresses these multiple levels of inference.
433 citations
TL;DR: The goal of this review was to systematize empirical research that was published in peer-reviewed English-language journals between 1995 and 2015 on the prevalence, predictors, and implications of adolescents’ use of pornography.
Abstract: The goal of this review was to systematize empirical research that was published in peer-reviewed English-language journals between 1995 and 2015 on the prevalence, predictors, and implications of adolescents' use of pornography. This research showed that adolescents use pornography, but prevalence rates varied greatly. Adolescents who used pornography more frequently were male, at a more advanced pubertal stage, sensation seekers, and had weak or troubled family relations. Pornography use was associated with more permissive sexual attitudes and tended to be linked with stronger gender-stereotypical sexual beliefs. It also seemed to be related to the occurrence of sexual intercourse, greater experience with casual sex behavior, and more sexual aggression, both in terms of perpetration and victimization. The findings of this review need to be seen against the background of various methodological and theoretical shortcomings, as well as several biases in the literature, which currently precludes internally valid causal conclusions about effects of pornography on adolescents.
319 citations
TL;DR: The results of this study support the directional association between screen time and child development and recommend encouraging family media plans, as well as managing screen time, to offset the potential consequences of excess use.
Abstract: Importance Excessive screen time is associated with delays in development; however, it is unclear if greater screen time predicts lower performance scores on developmental screening tests or if children with poor developmental performance receive added screen time as a way to modulate challenging behavior. Objective To assess the directional association between screen time and child development in a population of mothers and children. Design, Setting, and Participants This longitudinal cohort study used a 3-wave, cross-lagged panel model in 2441 mothers and children in Calgary, Alberta, Canada, drawn from the All Our Families study. Data were available when children were aged 24, 36, and 60 months. Data were collected between October 20, 2011, and October 6, 2016. Statistical analyses were conducted from July 31 to November 15, 2018. Exposures Media. Main Outcomes and Measures At age 24, 36, and 60 months, children’s screen-time behavior (total hours per week) and developmental outcomes (Ages and Stages Questionnaire, Third Edition) were assessed via maternal report. Results Of the 2441 children included in the analysis, 1227 (50.2%) were boys. A random-intercepts, cross-lagged panel model revealed that higher levels of screen time at 24 and 36 months were significantly associated with poorer performance on developmental screening tests at 36 months (β, −0.06; 95% CI, −0.10 to −0.01) and 60 months (β, −0.08; 95% CI, −0.13 to −0.02), respectively. These within-person (time-varying) associations statistically controlled for between-person (stable) differences. Conclusions and Relevance The results of this study support the directional association between screen time and child development. Recommendations include encouraging family media plans, as well as managing screen time, to offset the potential consequences of excess use.
272 citations
TL;DR: It is found that social media use is not a strong predictor of life satisfaction across the adolescent population and social media effects are nuanced, small at best, reciprocal over time, gender specific, and contingent on analytic methods.
Abstract: In this study, we used large-scale representative panel data to disentangle the between-person and within-person relations linking adolescent social media use and well-being. We found that social media use is not, in and of itself, a strong predictor of life satisfaction across the adolescent population. Instead, social media effects are nuanced, small at best, reciprocal over time, gender specific, and contingent on analytic methods.
248 citations
TL;DR: This work describes Latent change score modelling as a flexible statistical tool and provides accessible open source code and software examples to fit LCS models that can be readily formalized using key developmental questions.
238 citations
Cites background from "A critique of the cross-lagged pane..."
...Similarly, simple forms of the bivariate latent change score model can be rewritten as a special case of a cross-lagged panel model, namely the recently proposed random-intercept cross-lagged panel model (Hamaker et al., 2015), and the autoregressive cross-lagged factor model is equivalent to a latent change score model when slope factor scores are equivalent across individuals (Usami et al....
[...]
...Common procedures in developmental cognitive (neuro)science including cross-lagged panel models or simple regressions (on either raw or difference scores) can be considered special cases of LCSM’s, but without various benefits associated with SEM such as reduction of measurement error and incorporation of stable individual differences (Hamaker et al., 2015)....
[...]
References
More filters
Journal Article•
TL;DR: Copyright (©) 1999–2012 R Foundation for Statistical Computing; permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and permission notice are preserved on all copies.
Abstract: Copyright (©) 1999–2012 R Foundation for Statistical Computing. Permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and this permission notice are preserved on all copies. Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one. Permission is granted to copy and distribute translations of this manual into another language, under the above conditions for modified versions, except that this permission notice may be stated in a translation approved by the R Core Team.
272,030 citations
"A critique of the cross-lagged pane..." refers methods in this paper
...…seems to be an accepted sample size for a two-wave CLPM. Saving the parameter estimates in a separate file, which we then imported into R (R Core Team, 2012), we computed the standardized cross-lagged parameters (as Mplus does not allow for the computation of standardized parameters in case of…...
[...]
...We used a sample size of N 200, which seems to be an accepted sample size for a two-wave CLPM. Saving the parameter estimates in a separate file, which we then imported into R (R Core Team, 2012), we computed the standardized cross-lagged parameters (as Mplus does not allow for the computation of standardized parameters in case of Monte Carlo simulations)....
[...]
TL;DR: In this article, the cross spectrum between two variables can be decomposed into two parts, each relating to a single causal arm of a feedback situation, and measures of causal lag and causal strength can then be constructed.
Abstract: There occurs on some occasions a difficulty in deciding the direction of causality between two related variables and also whether or not feedback is occurring. Testable definitions of causality and feedback are proposed and illustrated by use of simple two-variable models. The important problem of apparent instantaneous causality is discussed and it is suggested that the problem often arises due to slowness in recording information or because a sufficiently wide class of possible causal variables has not been used. It can be shown that the cross spectrum between two variables can be decomposed into two parts, each relating to a single causal arm of a feedback situation. Measures of causal lag and causal strength can then be constructed. A generalisation of this result with the partial cross spectrum is suggested.
16,349 citations
01 Jan 2001
TL;DR: In this article, it is shown that the cross spectrum between two variables can be decomposed into two parts, each relating to a single causal arm of a feedback situation, and measures of causal lag and causal strength can then be constructed.
Abstract: There occurs on some occasions a difficulty in deciding the direction of causality between two related variables and also whether or not feedback is occurring. Testable definitions of causality and feedback are proposed and illustrated by use of simple two-variable models. The important problem of apparent instantaneous causality is discussed and it is suggested that the problem often arises due to slowness in recordhag information or because a sufficiently wide class of possible causal variables has not been used. It can be shown that the cross spectrum between two variables can be decomposed into two parts, each relating to a single causal arm of a feedback situation. Measures of causal lag and causal strength can then be constructed. A generalization of this result with the partial cross spectrum is suggested.The object of this paper is to throw light on the relationships between certain classes of econometric models involving feedback and the functions arising in spectral analysis, particularly the cross spectrum and the partial cross spectrum. Causality and feedback are here defined in an explicit and testable fashion. It is shown that in the two-variable case the feedback mechanism can be broken down into two causal relations and that the cross spectrum can be considered as the sum of two cross spectra, each closely connected with one of the causations. The next three sections of the paper briefly introduce those aspects of spectral methods, model building, and causality which are required later. Section IV presents the results for the two-variable case and Section V generalizes these results for three variables.
11,896 citations
"A critique of the cross-lagged pane..." refers background in this paper
...…Dwyer, 1983; Finkel, 1995; Heise, 1970), and diverse modeling strategies have been proposed to 1 While the omitted variable problem implies that we cannot make strong causal statements based on correlational data, it does not prohibit the use of the concept of Granger causality (Granger, 1969)....
[...]
...1 While the omitted variable problem implies that we cannot make strong causal statements based on correlational data, it does not prohibit the use of the concept of Granger causality (Granger, 1969)....
[...]
[...]
TL;DR: This paper provides a concise overview of time series analysis in the time and frequency domains with lots of references for further reading.
Abstract: Any series of observations ordered along a single dimension, such as time, may be thought of as a time series. The emphasis in time series analysis is on studying the dependence among observations at different points in time. What distinguishes time series analysis from general multivariate analysis is precisely the temporal order imposed on the observations. Many economic variables, such as GNP and its components, price indices, sales, and stock returns are observed over time. In addition to being interested in the contemporaneous relationships among such variables, we are often concerned with relationships between their current and past values, that is, relationships over time.
9,919 citations
"A critique of the cross-lagged pane..." refers background in this paper
...…were chosen to reflect diverse scenarios (e.g., no effects, a strong vs. a small effect etc.), but in all cases their values were smaller (in absolute value) than the autoregressive parameters, and they were chosen such that the bivariate process was covariance stationary (cf. Hamilton, 1994)....
[...]