Multilevel Modeling Methods Research Articles

Integrated Computational Materials Engineering: A Multi-Scale Approach

Computational materials science and engineering has a strong potential for developing robust simulation-based designs for optimizing materials, components, and engineering structures. Integrated computational materials engineering (ICME) provides practical solutions for engineering design tasks and has been recognized by the National Academy of Engineering (NAE Report 2008). The vital role of materials in design as well as the importance of computation in materials engineering necessitate understanding structure, properties, processing, and performance of materials and designed components. Thus, multiscale modeling has extended into solid mechanics, fluid mechanics, materials science, physics, mathematics, biology, and chemistry. ICME couples quantum scales with macroscales. It also allows the integration of materials processing and product performance for simulation-based manufacturing. These advances are analogous to transistor models that have led to complex logic operations and concurrent multilevel simulations. ICME can reduce product development time and costs by minimizing design cycles and facilitating cost-effective design optimization. These benefits have the potential to drive the utilization of multiscale modeling in various industrial sectors. The ‘‘integrated’’ part of ICME strongly emphasizes synergistic systems and interdisciplinary thinking accompanied by multiscale modeling methods. However, a key challenge is to link the various disciplines. Mathematics and statistics play a critical role and form a strong basis for modeling strategies. The terms ‘‘bottom-up’’ and ‘‘top-down’’ are of particular interest. The following characteristics are critical for these approaches: (I) length scales, (II) degrees of freedom and boundary conditions, (III) geometric and energy constraints, (IV) suitable numerical model implementation, (V) determining right sequence of calculations to be performed, (VI) determining variables and equations describing structure–property function, (VII) analyzing the cause–effect relationships using numerical methods, (VIII) verifying effects by experimental approaches before using the developed theory for next higher scales, and (IX) estimating the uncertainty in simulations and experiments to help develop a concrete multiscale modeling approach. This issue of JOM presents nine invited research and review articles. These articles discuss the importance of the ICME approach in refining conventional material and component processing methods as well as advanced materials processing. The article by Ghosh et al. discusses multiscale modeling of heterogeneous metals and alloys. They comprehensively address the problems related to hierarchical-concurrent multilevel or multiscale modeling methods for ductile fracture and nested dual-stage homogenization method for microstructure-based models for cast Al alloys. The two articles by Horstemeyer and colleagues detail the development and deployment of the modified embedded atom method (MEAM). The authors focus on the MATLAB-based MEAM potential calibration tool and interface with LAMMPS. The technique is used for studying aluminum system. To further define the ICME in ‘‘completeness,’’ the article by Liu et al. discusses case studies focused on three examples, where experiments are integrated with computation aspects. The modeling methods described include artificial neural networks, dislocation dynamics, and CALPHAD approach. Asadi et al. discuss the phase-field crystal model and multiscale analysis of solid–liquid structures. The article also compares various phase-field models for calculating solid–liquid interfaces. Bryan et al. demonstrate alloy design using the ICME-based Nitin Chopra is the Guest Editor for the ICME Committee of the TMS Materials Processing & Manufacturing Division, and coordinator of the topic Multi-Scale Modeling: Concurrent and Hierarchical Methods in this issue. JOM, Vol. 67, No. 1, 2015

Therapist effects and IAPT Psychological Wellbeing Practitioners (PWPs): A multilevel modelling and mixed methods analysis

The aim of this research was (a) to determine the extent of therapist effects in Psychological Wellbeing Practitioners (PWPs) delivering guided self-help in IAPT services and (b) to identify factors that defined effective PWP clinical practice. Using patient (N = 1122) anxiety and depression outcomes (PHQ-9 and GAD-7), the effectiveness of N = 21 PWPs across 6 service sites was examined using multi-level modelling. PWPs and their clinical supervisors were also interviewed and completed measures of ego strength, intuition and resilience. Therapist effects accounted for around 9 per cent of the variance in patient outcomes. One PWP had significantly better than average outcomes on both PHQ-9 and GAD-7 while 3 PWPs were significantly below average on the PHQ-9 and 2 were below average on the GAD-7. Computed PWP ranks identified quartile clusters of the most (N = 5) and least (N = 5) effective PWPs. More effective PWPs generated higher rates of reliable and clinically significant change and displayed greater resilience, organisational abilities, knowledge and confidence. Study weaknesses are identified and methodological considerations for future studies examining therapist effects in low intensity cognitive behaviour therapy are provided.

Identifying factors associated with the uptake of prevention of mother to child HIV transmission programme in Tigray region, Ethiopia: a multilevel modeling approach

BackgroundPrevention of mother to child HIV transmission (PMTCT) remains a challenge in low and middle-income countries. Determinants of utilization occur – and often interact - at both individual and community levels, but most studies do not address how determinants interact across levels. Multilevel models allow for the importance of both groups and individuals in understanding health outcomes and provide one way to link the traditionally distinct ecological- and individual-level studies. This study examined individual and community level determinants of mother and child receiving PMTCT services in Tigray region, Ethiopia.MethodsA multistage probability sampling method was used for this 2011 cross-sectional study of 220 HIV positive post-partum women attending child immunization services at 50 health facilities in 46 districts. In view of the nested nature of the data, we used multilevel modeling methods and assessed macro level random effects.ResultsSeventy nine percent of mothers and 55.7% of their children had received PMTCT services. Multivariate multilevel modeling found that mothers who delivered at a health facility were 18 times (AOR = 18.21; 95% CI 4.37,75.91) and children born at a health facility were 5 times (AOR = 4.77; 95% CI 1.21,18.83) more likely to receive PMTCT services, compared to mothers delivering at home. For every addition of one nurse per 1500 people, the likelihood of getting PMTCT services for a mother increases by 7.22 fold (AOR = 7.22; 95% CI 1.02,51.26), when other individual and community level factors were controlled simultaneously. In addition, district-level variation was low for mothers receiving PMTCT services (0.6% between districts) but higher for children (27.2% variation between districts).ConclusionsThis study, using a multilevel modeling approach, was able to identify factors operating at both individual and community levels that affect mothers and children getting PMTCT services. This may allow differentiating and accentuating approaches for different settings in Ethiopia. Increasing health facility delivery and HCT coverage could increase mother-child pairs who are getting PMTCT. Reducing the distance to health facility and increasing the number of nurses and laboratory technicians are also important variables to be considered by the government.

An Investigation of School-Level Factors Associated With Science Performance for Minority and Majority Francophone Students in Canada

AbstractMinority Francophone students in predominantly English-speaking Canadian provinces tend to perform lower on large-scale assessments of achievement than their Anglophone peers and majority Francophone students in Quebec. This study is the first to apply multilevel modeling methods to examine the extent to which school-level factors may be differentially associated with achievement for minority Francophone students compared to majority Francophone students. Forty-four percent of the variance in science achievement among Francophone schools was explained by the model. After taking into consideration gender as well as student- and school-level socioeconomic status, French-language schools in Ontario and New Brunswick had, on average, lower PISA 2006 science achievement scores compared to schools in Quebec. Findings indicate that differences in performance between linguistic minority and majority Francophone students could not be reduced to differences in socioeconomic status or to the school resource ...

The effects of community factors on school participation in Turkey: A multilevel analysis

Turkey, like many developing countries, is facing considerable problems in terms of low school attendance rates, late enrolment and early dropout of girls in particular. Numerous studies have already been conducted, both in Turkey and elsewhere, to determine the factors affecting school enrolment of boys and girls. Existing studies in Turkey, however, have focused extensively on the association between household-level factors and school participation, ignoring the role of the broader environment in which children live. Using a recent, large-scale and nationally representative data set, this paper investigates school participation at both primary and secondary levels in Turkey, giving specific attention to community-level factors. In taking into account socioeconomic context variables using the multilevel modelling method, this study contributes significantly to current school participation literature in Turkey. The author’s findings highlight the importance of community/context factors in explaining low school enrolment in Turkey. The results of the study can help policy makers develop a systematic understanding of the relationship between socioeconomic context and school participation, and enable them to make more appropriate decisions for improving school participation across the country.

Aggregating and Testing Intra-Individual Correlations: Methods and Comparisons

From a longitudinal study, we have repeatedly measured data from multiple individuals at multiple occasions. For each individual, the relation between 2 variables can be measured by the Pearson’s correlation. The question is how to aggregate the multiple correlations and conduct statistical inference on the aggregated intra-individual correlation. Several methods are proposed to aggregate and test intra-individual correlations: (a) a meta-analysis method based on Fisher’s Z transformed correlations, (b) a meta-analysis method based on the Pearson’s correlations, and (c) a multilevel modeling method using data standardized within each individual. The performance of the methods after bias corrections was compared using simulations with considering factors including numbers of individuals, numbers of time points, population effect sizes, and their distribution forms (homogeneous vs heterogeneous). The results from the simulation studies show that estimation biases were found using the meta-analytic methods and suggestions on when and how to correct biases were provided based on the simulation results. Furthermore, the performance of the 3 methods after necessary bias corrections was found to be comparable and reasonably good, indicating that all 3 methods worked for aggregating and testing intra-individual correlations. An empirical daily diary data set was then used to illustrate the applications of the 3 methods. The assumptions, advantages and disadvantages, and possible extensions of the 3 methods were discussed.

Statistical Power of Multilevel Modelling in Dental Caries Clinical Trials: A Simulation Study

Outcome data from dental caries clinical trials have a naturally hierarchical structure, with surfaces clustered within teeth, clustered within individuals. Data are often aggregated into the DMF index for each individual, losing tooth- and surface-specific information. If these data are to be analysed by tooth or surface, allowing exploration of effects of interventions on different teeth and surfaces, appropriate methods must be used to adjust for the clustered nature of the data. Multilevel modelling allows analysis of clustered data using individual observations without aggregating data, and has been little used in the field of dental caries. A simulation study was conducted to investigate the performance of multilevel modelling methods and standard caries increment analysis. Data sets were simulated from a three-level binomial distribution based on analysis of a caries clinical trial in Scottish adolescents, with varying sample sizes, treatment effects and random tooth level effects based on trials reported in Cochrane reviews of topical fluoride, and analysed to compare the power of multilevel models and traditional analysis. 40,500 data sets were simulated. Analysis showed that estimated power for the traditional caries increment method was similar to that for multilevel modelling, with more variation in smaller data sets. Multilevel modelling may not allow significant reductions in the number of participants required in a caries clinical trial, compared to the use of traditional analyses, but investigators interested in exploring the effect of their intervention in more detail may wish to consider the application of multilevel modelling to their clinical trial data.

OP08 Using Linked Administrative Data and Multilevel Modelling to Identify Targets for Interventions to Tackle Health Disparities

BackgroundLinked administrative data for whole populations, combined with multilevel modelling methods, offers a powerful approach to identifying targets for tackling health disparities. Two studies will be presented, examining disparities between...

A Meta-View of Multivariate Statistical Inference Methods in European Psychology Journals

We investigated the extent and nature of multivariate statistical inferential procedures used in eight European psychology journals covering a range of content (i.e., clinical, social, health, personality, organizational, developmental, educational, and cognitive). Multivariate methods included those found in popular texts that focused on prediction, group difference, and advanced modeling: multiple regression, logistic regression, analysis of covariance, multivariate analysis of variance, factor or principal component analysis, structural equation modeling, multilevel modeling, and other methods. Results revealed that an average of 57% of the articles from these eight journals involved multivariate analyses with a third using multiple regression, 17% using structural modeling, and the remaining methods collectively comprising about 50% of the analyses. The most frequently occurring inferential procedures involved prediction weights, dichotomous p values, figures with data, and significance tests with very few articles involving confidence intervals, statistical mediation, longitudinal analyses, power analysis, or meta-analysis. Contributions, limitations and future directions are discussed.

Mapping the Urban Debtscape: The Geography of Household Debt in Canadian Cities

Vulnerability resulting from debt is part and parcel of the risk society and a salient characteristic of current neoliberal times under financialized global capitalism. Rising indebtedness increases the susceptibility of homeowners to housing and labor market restructuring, and if the degree of leverage is very high, can threaten the solvency, living standards, and social stability of local communities. However, very little is understood regarding how levels of household indebtedness are spatially distributed within or across cities, and how private debt maps onto the geography of race, class, housing, urban form, and other social variables, especially outside of the United States. It remains unknown whether and how higher and unsustainable levels of indebtedness might be associated with urban growth, decline, suburbanization, gentrification, immigration, racialization, and/or greying. This article examines the spatial distribution of household debt in Canadian cities at multiple scales of analysis. It analyzes how levels of household debt relate to a number of key socio-demographic and housing variables from the census, including those related to changes occurring over the 2000s. It simultaneously models the geography of debt at the metropolitan and neighborhood scales using multi-level hierarchical linear modeling methods, and in doing so, it identifies some key drivers and correlates of household debt and the scales at which they operate. The article concludes by discussing the implications of the empirical findings for understanding the role of the emerging urban debtscape in the restructuring of the social geography of the city.

Regression Analysis of Country Effects Using Multilevel Data: A Cautionary Tale

Cross-national differences in outcomes are often analysed using regression analysis of multilevel country datasets, examples of which include the ECHP, ESS, EU-SILC, EVS, ISSP, and SHARE. We review the regression methods applicable to this data structure, pointing out problems with the assessment of country-level factors that appear not to be widely appreciated, and illustrate our arguments using Monte-Carlo simulations and analysis of women’s employment probabilities and work hours using EU SILC data. With large sample sizes of individuals within each country but a small number of countries, analysts can reliably estimate individual-level effects within each country but estimates of parameters summarising country effects are likely to be unreliable. Multilevel (hierarchical) modelling methods are commonly used in this context but they are no panacea.

Incidence of renal replacement therapy in Veneto for 2008, 2009 and 2010

This section of the report of the Veneto Dialysis and Transplantation Registry (VDTR) provides data on the incidence of patients receiving renal replacement therapy (RRT) in the region from 2008 to 2010. Its purpose is to provide health authorities with the information they need to plan the delivery of RRT in Veneto. Data were obtained from the VDTR, defining incident patients according to the recommendations of the Italian Dialysis and Transplantation Registry. The incidence rate was calculated per million population (pmp). Variability by province and treatment center was studied by applying multilevel modeling methods. An age-period-cohort model was used to forecast the incidence rate of RRT over the years to come. The incidence of patients on RRT was 114.23 pmp in 2008, 120.15 pmp in 2009 and 107.08 pmp in 2010. The patients' median age at the time of starting RRT was 70.5 in 2008, 68.7 in 2009 and 69.5 in 2010. During these 3 years, 66.3% of patients were male, and 33.7% were female. Incidence rates were not uniformly distributed between the provinces in the region, but were significantly higher in 2. The incidence rate of patients needing RRT seems likely to remain stable in the future, until 2015 at least. Renal vascular disease was the primary cause of end-stage renal disease (ESRD), followed closely by diabetes, while the proportion due to primary glomerulonephritis has gradually decreased. Initial dialysis modality was hemodialysis (HD) for 78% of patients, while about 20% started RRT on peritoneal dialysis (PD), and a negligible proportion had a preemptive kidney transplantation. About 35% patients began dialysis with a temporary vascular catheter; this percentage remained fairly constant until 2010. The incidence of RRT in Veneto is one of the lowest in Italy and remained substantially stable over the period 1998-2010, despite the population of patients with ESRD becoming older and more severely ill. This finding could mean a heavier burden on the welfare system in the future.

Multilevel Modeling of the Performance Variance

Focusing on the identification of the role played by Industry on the relations between Corporate Strategic Factors and Performance, the hierarchical multilevel modeling method was adopted when measuring and analyzing the relations between the variables that comprise each level of analysis. The adequacy of the multilevel perspective to the study of the proposed relations was identified and the relative importance analysis point out to the lower relevance of industry as a moderator of the effects of corporate strategic factors on performance, when the latter was measured by means of return on assets, and that industry don‟t moderates the relations between corporate strategic factors and Tobin‟s Q. The main conclusions of the research are that the organizations choices in terms of corporate strategy presents a considerable influence and plays a key role on the determination of performance level, but that industry should be considered when analyzing the performance variation despite its role as a moderator or not of the relations between corporate strategic factors and performance.

MULTILEVEL MODELING OF THE PERFORMANCE VARIANCE

Focusing on the identification of the role played by Industry on the relations between Corporate Strategic Factors and Performance, the hierarchical multilevel modeling method was adopted when measuring and analyzing the relations between the variables that comprise each level of analysis. The adequacy of the multilevel perspective to the study of the proposed relations was identified and the relative importance analysis point out to the lower relevance of industry as a moderator of the effects of corporate strategic factors on performance, when the latter was measured by means of return on assets, and that industry don‟t moderates the relations between corporate strategic factors and Tobin‟s Q. The main conclusions of the research are that the organizations choices in terms of corporate strategy presents a considerable influence and plays a key role on the determination of performance level, but that industry should be considered when analyzing the performance variation despite its role as a moderator or not of the relations between corporate strategic factors and performance.

An introduction to generalized linear models—third edition

- Introduces GLMs in a way that enables readers to understand the unifying structure that underpins them. - Discusses common concepts and principles of advanced GLMs, including nominal and ordinal regression, survival analysis, and longitudinal analysis. - Connects Bayesian analysis and MCMC methods to fit GLMs. - Contains numerous examples from business, medicine, engineering, and the social sciences. - Provides the example code for R, Stata, and WinBUGS to encourage implementation of the methods. - Offers the data sets and solutions to the exercises online. Continuing to emphasize numerical and graphical methods, An Introduction to Generalized Linear Models, Third Edition provides a cohesive framework for statistical modeling. This new edition of a bestseller has been updated with Stata, R, and WinBUGS code as well as three new chapters on Bayesian analysis. Like its predecessor, this edition presents the theoretical background of generalized linear models (GLMs) before focusing on methods for analyzing particular kinds of data. It covers normal, Poisson, and binomial distributions; linear regression models; classical estimation and model fitting methods; and frequentist methods of statistical inference. After forming this foundation, the authors explore multiple linear regression, analysis of variance (ANOVA), logistic regression, log-linear models, survival analysis, multilevel modeling, Bayesian models, and Markov chain Monte Carlo (MCMC) methods. Using popular statistical software programs, this concise and accessible text illustrates practical approaches to estimation, model fitting, and model comparisons. It includes examples and exercises with complete data sets for nearly all the models covered.

A Framework of Digital Factory System Using Multi-Resolution Simulation

Digital Factory (DF) technology is one of the most important part of manufacturer enterprise information systems, its core technology is modeling & simulation. This paper introduced the digital factory concept and its functions with multi-level multi-resolution modeling and simulation method; we established a new framework of the digital factory technology. It combines with the features of DF and simulation technology which expands the development of DF and application of simulation technology. A case was taken to show the feasibility of multi-level modeling & multi-resolution simulation technology attaching to the DF technology.

Research on Multi-Level Modeling Method for Aircraft Assembly Line

The process time of aircraft assembly line is easily influenced by the efficiency of workers, the supply circumstance of components and the assembly quality. While the aircraft assembly line is hardly to reconfigure after it is in use, in order to optimize the assembly line before its application, a new object-oriented modeling method based on hierarchy and stochastic petri net has been put forward in this paper. The behavior and the structural characteristics of the model have been analyzed according to the solution of linear invariant. Finally, the model of aircraft wing assembly line has been established as an example, which proves the feasibility of the modeling method.

Improving Statistical Inference with Clustered Data

Political science data often contain grouped observations, which produces unobserved "cluster effects" in statistical models. Typical solutions include (1) ignoring the impact on coefficients and only adjusting the standard errors of generalized linear models (GLM) or (2) addressing clustering in coefficient estimation while relying on a parametric assumption for the cluster effects and/or a large number of clusters for standard errors. I show that both approaches are problematic for inference. Through simulation I demonstrate that multilevel modeling (MLM) and generalized estimating equations (GEE) produce more efficient coefficients than does GLM. Next, I show that commonly-used MLM and GEE standard error methods can be biased downward, while bootstrapping by resampling clusters (BCSE) performs better, even with a misspecified error distribution and/or few clusters. I recommend the use of MLM or GEE to estimate coefficients and BCSE to estimate uncertainty, and show that this approach can produce divergent conclusions in applied research.

A new rational-based optimal design strategy of ship structure based on multi-level analysis and super-element modeling method

A new multi-level analysis method of introducing the super-element modeling method, derived from the multi-level analysis method first proposed by O. F. Hughes, has been proposed in this paper to solve the problem of high time cost in adopting a rational-based optimal design method for ship structural design. Furthermore, the method was verified by its effective application in optimization of the mid-ship section of a container ship. A full 3-D FEM model of a ship, suffering static and quasi-static loads, was used as the analyzing object for evaluating the structural performance of the mid-ship module, including static strength and buckling performance. Research results reveal that this new method could substantially reduce the computational cost of the rational-based optimization problem without decreasing its accuracy, which increases the feasibility and economic efficiency of using a rational-based optimal design method in ship structural design.

Accounting for the Heterogeneity in Inter-generational Links in Educational Attainment Across Europe

We use data drawn from the European Social Survey to investigate the extent of inter-generational links in educational attainment across twenty-four European countries. We find that there is a lot of cross-country variation in the observed patterns especially when gender distinction is made at both the parents’ and the children’s generations. To account for the variation in the estimates obtained in separate country regressions, we propose the use of simple educational inequality measures as country-level control variables. While the variation in the children’s educational attainment turns out to be positively associated with the level of inter-generational links, the opposite is the case for the variation in the parents’ attainment. The introduction of country-level variables into the analysis brings about the estimation of the econometric models on the pooled sample of all countries and the use of multilevel modeling methods which we find to perform better than least squares estimation.