Naslov
Testing mediation in cross-legged panel designs

Autori
Buško, Vesna

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
18. Dani Ramira i Zorana Bujasa: Program i sažeci priopćenja / Keresteš, Gordana ; Brković, Irma ; Butković, Ana - Zagreb : Odsjek za psihologiju Filozofskog fakulteta Sveučilišta u Zagrebu, 2007, 46-46

ISBN
978-953-175-302-9

Skup
18. Dani Ramira i Zorana Bujasa

Mjesto i datum
Zagreb, Hrvatska, 13.12.2007. - 15.12.2007

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Domaća recenzija

Ključne riječi
structural equation modeling; mediation testing; panel design

Sažetak
Mediation has been defined as the process or mechanism by which an intervening mediator variable (M) carries the eff ect of an independent variable (X) to a dependent variable (Y). Methodological literature gives an extensive account of the statistical approaches to the analysis mediation and different methods for assessing the magnitude and significance of indirect effects. Mediation effects are typically quantified as the product of the path coefficient linking X to mediator M and the coefficient linking M to Y while controlling for X. Differences among the methods pertain mainly to the problem of how to assess the significance of this quantity. As mediation is by definition a causal process, testing indirect effects hypothesis against cross-sectional data may be misleading. Panel designs off er means to correct for the related problems. A structural equation model used to demonstrate this strategy for testing mediation was specifi ed within a 2-wave cross-lagged panel design, with stationarity assumed on theoretical grounds. The empirical data used for this purpose were based on the follow-up study on the processes of coping with low-control stress. The results are compared with those obtained by structural equation analyses performed within diff erent methodological framework but tested against the same data sets. These include structural equation models specified in line with the assumptions of latent state-trait theory (LST ; Steyer et al., 1992, 1999), and using the true change modeling procedures.

Izvorni jezik
Engleski

Znanstvena područja
Psihologija

Napomena
Rad je prezentiran na "3rd Psychometric Symposium in honour of Professor Alija Kulenović" održanom u sklopu 18. Dana Ramira i Zorana Bujasa, Zagreb, prosinac 2007.

POVEZANOST RADA