The Impact of Misspecifying Class-Specific Residual Variances in Growth Mixture Models

Enders, Craig K.; Tofighi, Davood

doi:10.1080/10705510701758281

Cited by 114 publications

(92 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A corrected version of the BIC index (sample size adjusted BIC-SA-BIC) was used in the present study. This modification has been proposed as the uncorrected BIC may place too high a penalty on model complexity (see: Enders and Tofighi 2008;Tofghi and Enders 2007). 2 Prior to performing the analyses below, a series of ancillary analyses evaluated the adequacy of the present sample to inferentially support the tested hypotheses.…”

Section: Confirmatory Factor Analysis Comparing Dsm-iv-tr and Dsm-5 Amentioning

confidence: 98%

Comparing Diagnostic Outcomes of Autism Spectrum Disorder Using DSM-IV-TR and DSM-5 Criteria

Harstad

Fogler

Sideridis

et al. 2014

J Autism Dev Disord

View full text Add to dashboard Cite

Controversy exists regarding the DSM-5 criteria for ASD. This study tested the psychometric properties of the DSM-5 model and determined how well it performed across different gender, IQ, and DSM-IV-TR sub-type, using clinically collected data on 227 subjects (median age = 3.95 years, majority had IQ > 70). DSM-5 was psychometrically superior to the DSM-IV-TR model (Comparative Fit Index of 0.970 vs 0.879, respectively). Measurement invariance revealed good model fit across gender and IQ. Younger children tended to meet fewer diagnostic criteria. Those with autistic disorder were more likely to meet social communication and repetitive behaviors criteria (p < .001) than those with PDD-NOS. DSM-5 is a robust model but will identify a different, albeit overlapping population of individuals compared to DSM-IV-TR.

show abstract

Section: Confirmatory Factor Analysis Comparing Dsm-iv-tr and Dsm-5 Amentioning

confidence: 98%

Comparing Diagnostic Outcomes of Autism Spectrum Disorder Using DSM-IV-TR and DSM-5 Criteria

Harstad

Fogler

Sideridis

et al. 2014

J Autism Dev Disord

View full text Add to dashboard Cite

show abstract

“…Previous studies had already shown that sample size (e.g., Enders & Tofighi, 2008;Li, & Hser, 2011;Liu & Hancock, 2014;Lubke & Muthén, 2007;Peugh & Fan, 2012Tofighi & Enders, 2007) and having an equal mixing ratio (e.g., Enders & Tofighi, 2008;Nylund et al 2007;Peugh & Fan, 2012Tofighi & Enders, 2007) had a positive impact on the class enumeration and parameter recovery of GMMs. However, this study was the first to systematically investigate the impact of the residual design factors.…”

Section: Manipulated Factorsmentioning

confidence: 99%

“…Furthermore, although Tolvanen (2008) considered the effects of the R 2 conditions, he did not consider the fact that the residual structure could be either heteroscedastic or homoscedastic across time points. The homoscedasticity of the residuals is a common assumption in LGM (e.g., Cheong, 2011;Fan & Fan, 2005), which has been shown to result in biases in GMM when the population model is heteroscedastic (Enders & Tofighi, 2008).…”

Section: Growth Mixture Models With Tvcsmentioning

confidence: 99%

“…Another question that remains unanswered is related to the impact of the residual structure on the class enumeration accuracy of GMMs (e.g., Enders & Tofighi, 2008;Tolvanen, 2008)-another design factor that may prove critical in the presence of TVCs and was not investigated by Peugh and Fan (2015). Indeed, the influence of TVCs on the repeated measures is directly, and inversely, related to the size of the residuals (i.e., or directly related to the R 2 of the repeated measures), and is also intimately related to the structure of these residuals (homoscedastic, heteroscedastic, class-invariant, class-varying; see the online supplements).…”

Section: Growth Mixture Models With Tvcsmentioning

confidence: 99%

“…Four measurement points are often used in simulation studies of GMMs (Enders & Tofighi, 2008;Liu & Hancock, 2014;Nylund et al, 2007;Peugh & Fan, 2015; Residual variances Homoscedastic condition…”

Section: Manipulated Factorsmentioning

confidence: 99%

See 2 more Smart Citations

Performance of growth mixture models in the presence of time-varying covariates

Diallo

Morin

2016

Behav Res

View full text Add to dashboard Cite

Growth mixture modeling is often used to identify unobserved heterogeneity in populations. Despite the usefulness of growth mixture modeling in practice, little is known about the performance of this data analysis technique in the presence of time-varying covariates. In the present simulation study, we examined the impacts of five design factors: the proportion of the total variance of the outcome explained by the time-varying covariates, the number of time points, the error structure, the sample size, and the mixing ratio. More precisely, we examined the impact of these factors on the accuracy of parameter and standard error estimates, as well as on the class enumeration accuracy. Our results showed that the consistent Akaike information criterion (CAIC), the sample-size-adjusted CAIC (SCAIC), the Bayesian information criterion (BIC), and the integrated completed likelihood criterion (ICL-BIC) proved to be highly reliable indicators of the true number of latent classes in the data, across design conditions, and that the sample-size-adjusted BIC (SBIC) also proved quite accurate, especially in larger samples. In contrast, the Akaike information criterion (AIC), the entropy, the normalized entropy criterion (NEC), and the classification likelihood criterion (CLC) proved to be unreliable indicators of the true number of latent classes in the data. Our results also showed that substantial biases in the parameter and standard error estimates tended to be associated with growth mixture models that included only four time points.

show abstract

The growth of gossip: Socialization of relational aggression from adolescence to emerging adulthood

Coyne

Swit

Stockdale

et al. 2020

Aggressive Behavior

View full text Add to dashboard Cite

The extant literature suggests that relational aggression appears in early childhood, and gradually increases throughout adolescence. However, very little research has examined the growth of relational aggression from adolescence to emerging adulthood. In addition, research generally examines socializing factors of relational aggression, such as parenting, peers, siblings, or media in isolation. Accordingly, the aim of the current study was to examine these socializing factors conjunctively as predictors of the growth of relational aggression over time. Participants consisted of 500 adolescents who completed several questionnaires over a 7-year period (between ages 14-20 on average). Results revealed that the vast majority of individuals (88%) showed low levels of relational aggression that decreased over time. Conversely, a small proportion of individuals (12%) had high, increasing levels of relational aggression between adolescence and emerging adulthood. High levels of maternal psychological control, sibling hostility, and relational aggression in the media at the initial time point all predicted being in the high and increasing group.

show abstract

The Impact of Misspecifying Class-Specific Residual Variances in Growth Mixture Models

Cited by 114 publications

References 21 publications

Comparing Diagnostic Outcomes of Autism Spectrum Disorder Using DSM-IV-TR and DSM-5 Criteria

Comparing Diagnostic Outcomes of Autism Spectrum Disorder Using DSM-IV-TR and DSM-5 Criteria

Performance of growth mixture models in the presence of time-varying covariates

The growth of gossip: Socialization of relational aggression from adolescence to emerging adulthood

Contact Info

Product

Resources

About