An Application of Multilevel Model Prediction to NELS:88

Afshartous, David; Leeuw, Jan de

doi:10.2333/bhmk.31.43

Cited by 5 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Multi-level modeling has been widely used in research in education, demography, psychology, sociology, and agriculture. [13][14][15] To our knowledge, application of this analytical tool in studies of infectious disease has been limited. A few studies have been carried out on schistosomiasis infection in Africa, [16][17][18] mostly used Bayesian analysis or Bayesian spatial analysis of schistosomiasis infection in school children, including one study on the relationships between anemia and parasitic infections (including schistosomiasis) in Kenyan school children.…”

Section: Discussionmentioning

confidence: 99%

A multi-level analysis of risk factors for Schistosoma japonicum infection in China

Yang¹,

Zhao²,

Li³

et al. 2009

International Journal of Infectious Diseases

View full text Add to dashboard Cite

The aim of this study was to explore the risk factors of schistosomiasis japonica in China, using a hierarchical multi-level model with individuals nested within villages. Methods: A cross-sectional survey of schistosomiasis japonica was conducted in 16 villages in the Chinese province of Hunan. A multi-level modeling technique (HLM version 6.04) was used to assess risk factors of schistosomiasis. The results from this multi-level model were compared with those from a conventional single-level logistic regression model. Results: A total of 10 245 individuals were enrolled in this study, of whom about 4.1% were infected with Schistosoma japonicum. In the multi-level model analysis, individual level variables such as gender, age, and occupation, and village level variables such as type of S. japonicum endemic area, drinking water source, sewage treatment, June temperature, and April rainfall were associated with schistosomiasis japonica infection. Conventional single-level logistic regression analysis selected more independent variables, and had narrower confidence intervals around the corresponding regression coefficients. In particular, per capita income, precipitation in October, and density of infected snails were statistically significant in the conventional singlelevel logistic regression analysis but not in the multi-level model.

show abstract

Section: Discussionmentioning

confidence: 99%

A multi-level analysis of risk factors for Schistosoma japonicum infection in China

Yang¹,

Zhao²,

Li³

et al. 2009

International Journal of Infectious Diseases

View full text Add to dashboard Cite

show abstract

“…Consider the National Educational Longitudinal Study data set that was used by Afshartous and de Leeuw (2004), where the base year sample from the 1988 study is used (National Center for Educational Statistics, 2006). The base year sample consists of 24599 eighth‐grade students, distributed among 1052 schools across the USA.…”

Section: Applicationsmentioning

confidence: 99%

Avoiding ‘Data Snooping’ in Multilevel and Mixed Effects Models

Afshartous

Wolf

2007

Journal of the Royal Statistical Society Series A: Statistics in Society

Self Cite

View full text Add to dashboard Cite

Summary. Multilevel or mixed effects models are commonly applied to hierarchical data. The level 2 residuals, which are otherwise known as random effects, are often of both substantive and diagnostic interest. Substantively, they are frequently used for institutional comparisons or rankings. Diagnostically, they are used to assess the model assumptions at the group level. Inference on the level 2 residuals, however, typically does not account for 'data snooping', i.e. for the harmful effects of carrying out a multitude of hypothesis tests at the same time. We provide a very general framework that encompasses both of the following inference problems: inference on the 'absolute' level 2 residuals to determine which are significantly different from 0, and inference on any prespecified number of pairwise comparisons. Thus, the user has the choice of testing the comparisons of interest. As our methods are flexible with respect to the estimation method that is invoked, the user may choose the desired estimation method accordingly. We demonstrate the methods with the London education authority data, the wafer data and the National Educational Longitudinal Study data.

show abstract