On the utilization of sample weights in latent variable models

Kaplan, David; Ferguson, Aaron J.

doi:10.1080/10705519909540138

Cited by 66 publications

(53 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Statisticians have demonstrated that, under these conditions, population variables, standard errors, and variances are biased when estimated from sample data. 26 These potential type I errors can be ameliorated by calculating sampling weights with the Taylor series approximation and incorporating these sampling weights into subsequent analyses. 27 Missing data imputation was unnecessary, because the study Web site prevented participants with missing responses on a given measure from proceeding to subsequent measures in the protocol.…”

Section: Methodsmentioning

confidence: 99%

Low Burnout and High Engagement Levels in Athletic Trainers: Results of a Nationwide Random Sample

Giacobbi

2009

Journal of Athletic Training

View full text Add to dashboard Cite

Objective: To assess the prevalence of occupational burnout, engagement, and somatic health complaints and the associations among these constructs experienced by certified athletic trainers (ATs). A secondary objective was to examine differences between men and women, those with more versus less postcertification experience, and ATs who worked in different occupational settings.Design: Survey-based, stratified, proportionate random sample of full-time ATs.Setting: Data were collected online from ATs employed full time in the college or university, secondary school or youth, and industrial or clinical settings.Patients or Other Participants: Of 3998 invited ATs, 934 replied, for a response rate of 24%. The mean age of the participants was 33.84 6 8.29 years, and each AT worked with an average of 90 clients.Main Outcome Measure(s): The dependent variables were perceived stress, occupational burnout, engagement, and somatic health complaints.Results: Although 17.2% of participants were in the most advanced stages of burnout, low levels of burnout and high degrees of occupational engagement were observed. Women ATs and those working in the college or university settings scored higher for burnout, whereas men and those in the clinical or industrial settings scored higher for engagement. Women also reported significantly more somatic health complaints than men did.Conclusions: The overall prevalence of burnout in ATs was relatively small, but sex and occupational-setting differences were observed and deserve greater scrutiny.Key Words: stress, health, athletic training careers, psychology Key Points N Female athletic trainers and athletic trainers working in college or university settings displayed more signs of burnout. N Occupational stress was positively related to various elements of burnout and somatic health complaints but negatively related to occupational engagement.N Compared with other health care providers, athletic trainers were relatively less burned out and exhibited higher levels of engagement.

show abstract

Section: Methodsmentioning

confidence: 99%

Low Burnout and High Engagement Levels in Athletic Trainers: Results of a Nationwide Random Sample

Giacobbi

2009

Journal of Athletic Training

View full text Add to dashboard Cite

show abstract

“…This study clearly demonstrated that this method overestimates the chi-square value and underestimates the asymptotic covariance of the parameter estimates, resulting in low confidence interval coverage. The problems of this method were also demonstrated in Kaplan and Ferguson (1999). In a somewhat different context Stapleton (2002) demonstrated this as well.…”

Section: Comparison Among Mplus Mlwin and Hlm/sas Proc Mixedmentioning

confidence: 74%

“…One way to avoid this problem is to assume that the population is not infinite but a finite large population so that p > 0 and as a whole the large finite population would be numerically equivalent to an infinite population. Indeed many simulation studies in the complex sampling literature are designed that way; for example, Kaplan and Ferguson (1999) and Pfeffermann, Skinner, Holmes, Goldstein, and Rasbash (1998). Alternatively we can assume that the population is infinite and that p only represents a relative fre-…”

Section: Definitions and Interpretationmentioning

confidence: 99%

Sampling Weights in Latent Variable Modeling

Asparouhov¹

2005

Structural Equation Modeling: A Multidisciplinary Journal

503

313

View full text Add to dashboard Cite

This article reviews several basic statistical tools needed for modeling data with sampling weights that are implemented in Mplus Version 3. These tools are illustrated in simulation studies for several latent variable models including factor analysis with continuous and categorical indicators, latent class analysis, and growth models. The pseudomaximum likelihood estimation method is reviewed and illustrated with stratified cluster sampling. Additionally, the weighted least squares method for estimating structural equation models with categorical and continuous outcomes implemented in Mplus extended to incorporate sampling weights is also illustrated. The performance of several chi-square tests under unequal probability sampling is evaluated. Simulation studies compare the methods used in several statistical packages such as Mplus, HLM, SAS Proc Mixed, MLwiN, and the weighted sample statistics method used in other software packages.

show abstract

“…Importantly, ML-SEM, which is a hybrid model of conventional structural equation modeling and hierarchical linear modeling, would prevent biased structural regression coefficients (Muthén & Satorra, 1989). Compared to conventional and single-level SEM, ML-SEM can provide more accurate and unbiased estimates of population parameters because it takes into account hierarchically nested systems that most educational datasets have (Muthén & Satorra, 1989;Muthén & Muthén, 1998;Kaplan & Ferguson, 1999). As the Educational Longitudinal Study of 2002 (which the current study used) has a nested structure, ML-SEM would be necessary to report unbiased results.…”

Section: Methodsmentioning

confidence: 99%

Determining the Effects of Pre-College STEM Contexts on STEM Major Choices in 4-year Postsecondary Institutions Using Multilevel Structural Equation Modeling

Lee¹

2013

Journal of Pre-College Engineering Education Research (J-PEER)

View full text Add to dashboard Cite

Many STEM studies have focused on traditional learning contexts, such as math-and science-related learning factors, as pre-college learning predictors for STEM major choices in colleges. Few studies have considered a progressive learning activity embedded within STEM contexts. This study chose computer-based learning activities in K-12 math classrooms as a major pre-college learning predictor for STEM major choices. Using a nationally represented sample drawn from the Educational Longitudinal Study of 2002/06, the purpose of this study was two-fold: (a) to investigate the influence of computer-based learning activities in math classrooms on STEM major choices in 4-year post-secondary institutions and (b) to analyze the extent to which math teacher motivation affects math performance and math self-efficacy across schools, which plays a vital role in students' STEM major selection. The multilevel structural equation modeling revealed several findings. First, computer-based learning activities had a greater positive effect on math self-efficacy that significantly predicted the increase in the proportion of students' STEM major choices, compared to the individual-and lecture-based learning activities. Second, a positive relationship between individual-based learning activities and math performance emerged, which was associated with the high proportion of students' STEM major choices. Third, at the high school level, math teacher motivation positively influenced math performance. These results suggest that integrating STEM contexts into traditional learning activities in math curriculum at the K-12 level would increase students' interests in studying STEM disciplines. Equally important is enhancing math teachers' motivation, given the fact that these teachers design and implement the math curriculum.

show abstract

On the utilization of sample weights in latent variable models

Cited by 66 publications

References 13 publications

Low Burnout and High Engagement Levels in Athletic Trainers: Results of a Nationwide Random Sample

Low Burnout and High Engagement Levels in Athletic Trainers: Results of a Nationwide Random Sample

Sampling Weights in Latent Variable Modeling

Determining the Effects of Pre-College STEM Contexts on STEM Major Choices in 4-year Postsecondary Institutions Using Multilevel Structural Equation Modeling

Contact Info

Product

Resources

About