Assessing Model Selection Uncertainty Using a Bootstrap Approach: An Update

Lubke, Gitta H.; Campbell, Ian; McArtor, Dan; Miller, Patrick J. O.; Luningham, Justin M.; Berg, Stéphanie Martine van den

doi:10.1080/10705511.2016.1252265

Cited by 33 publications

(28 citation statements)

References 32 publications

(50 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the factorial variables (hypothesis, software package and multiple testing method), this was not possible because there is not a single coefficient for the factor; in addition, for software package and multiple testing methods, some bootstrap samples did not contain all values of the factor. For these variables we instead performed model comparison between the full model and a reduced model excluding each factor, and computed the proportion of times the full model was selected on the basis of the model selection criterion (using both Bayesian information criterion and Akaike information criterion) being numerically lower in the full model 27 .…”

Section: Factors Related To Analytical Variabilitymentioning

confidence: 99%

Variability in the analysis of a single neuroimaging dataset by many teams

Botvinik‐Nezer

Holzmeister

Camerer

et al. 2020

Nature

700

447

View full text Add to dashboard Cite

Section: Factors Related To Analytical Variabilitymentioning

confidence: 99%

Variability in the analysis of a single neuroimaging dataset by many teams

Botvinik‐Nezer

Holzmeister

Camerer

et al. 2020

Nature

700

447

View full text Add to dashboard Cite

“…As both criteria led to the same results (with only slight differences in decimal places), we will only report the results based on ΔBIC here (see Table 5). Please note that the BIC tends to prefer models with less parameters, especially in small samples, and that fit indices are probabilistic, and not absolute, criteria (Lubke et al, 2017). For openness, an accentuated long-term effect fit comparably better than a reversed effect in the multivariate (ΔBICreversed-accentuated, multivariate = 5.51), but not in the univariate model (ΔBIC reversed-accentuated, univariate = 1.89).…”

Section: Analyses Across T1-t2-t3-t4mentioning

confidence: 99%

Do Sojourn Effects on Personality Trait Changes Last? A Five–Year Longitudinal Study

et al. 2020

View full text Add to dashboard Cite

This study examined sojourners' long-term personality trait changes over five years, extending previous research on immediate sojourn effects. A sample of German students (N = 1095) was surveyed thrice (T1-T3) over the course of an academic year. Sojourners (n = 498) lived abroad shortly after T1 for one or two semesters; stayers (n = 597) remained in their home country. Five years after T1, we surveyed the same participants (n = 441, 40.3%) again (T4). Beyond substantial selection effects, latent neighbour-change models revealed that small differences between sojourners' and stayers' openness, agreeableness, and neuroticism changes occurred early after sojourn-induced contextual change. Model estimates suggested sustained sojourn effects on openness and neuroticism changes thereafter and a reversed effect on agreeableness change after return. Because of reduced power and low accuracy at T4, these estimates were not statistically significant. Based on model comparison analyses, however, we could rule out reversed effects for openness and accentuated effects for agreeableness and neuroticism as least likely. Moreover, separating short-term and long-term sojourners revealed no substantial differences, but recurring sojourn experiences tended to play a role in sustaining differences. We discuss implications for future studies on patterns of sojourn effects on personality trait changes.

show abstract

“…Hastie et al (2001) claimed that the SRM method performs poorly and suggested that AIC results in a superior performance. Lubke et al (2017) performed a simulation study for selecting a model via a bootstrap approach. In addition, Vrieze (2012) addressed the difference between the statistics AIC and BIC focusing on latent variable models.…”

Section: Introductionmentioning

confidence: 99%

Heteroscedasticity and Model Selection via Partitioning in Fisheries Data

Marzjarani¹

2018

IJSP

View full text Add to dashboard Cite

Selecting a proper model for a data set is a challenging task. In this article, an attempt was made to answer and to find a suitable model for a given data set. A general linear model (GLM) was introduced along with three different methods for estimating the parameters of the model. The three estimation methods considered in this paper were ordinary least squares (OLS), generalized least squares (GLS), and feasible generalized least squares (FGLS). In the case of GLS, two different weights were selected for improving the severity of heteroscedasticity and the proper weight (s) was deployed. The third weight was selected through the application of FGLS. Analyses showed that only two of the three weights including the FGLS were effective in improving or reducing the severity of heteroscedasticity. In addition, each data set was divided into Training, Validation, and Testing producing a more reliable set of estimates for the parameters in the model. Partitioning data is a relatively new approach is statistics borrowed from the field of machine learning. Stepwise and forward selection methods along with a number of statistics including the average square error testing (ASE), Adj. R-Sq, AIC, AICC, and ASE validate along with proper hierarchies were deployed to select a more appropriate model(s) for a given data set. Furthermore, the response variable in both data files was transformed using the Box-Cox method to meet the assumption of normality. Analysis showed that the logarithmic transformation solved this issue in a satisfactory manner. Since the issues of heteroscedasticity, model selection, and partitioning of data have not been addressed in fisheries, for introduction and demonstration purposes only, the 2015 and 2016 shrimp data in the Gulf of Mexico (GOM) were selected and the above methods were applied to these data sets. At the conclusion, some variations of the GLM were identified as possible leading candidates for the above data sets.

show abstract

Assessing Model Selection Uncertainty Using a Bootstrap Approach: An Update

Cited by 33 publications

References 32 publications

Variability in the analysis of a single neuroimaging dataset by many teams

Variability in the analysis of a single neuroimaging dataset by many teams

Do Sojourn Effects on Personality Trait Changes Last? A Five–Year Longitudinal Study

Heteroscedasticity and Model Selection via Partitioning in Fisheries Data

Contact Info

Product

Resources

About