Home About us Contact
|
Home About us Contact | ||
|
|
||
Multilevel Models (multilevel + models)
Selected AbstractsAccounting for Aberrant Test Response Patterns Using Multilevel ModelsJOURNAL OF EDUCATIONAL MEASUREMENT, Issue 3 2007Alexandra Petridou Hypotheses about aberrant test-response behavior and hence invalid person-measurement have hitherto included factors like ability, gender, language, test-anxiety, and motivation, but these have not previously been collectively investigated with real data, or with multilevel models. This study analyzes the effect of these factors on person aberrance using a real mathematics assessment data set under the framework of a two-level (person and classroom) hierarchical model. The results suggest that higher-scoring pupils, and, to a lesser extent, second-language learners are significantly more often aberrant. But more importantly, we find that the classroom makes a significant contribution to person aberrance and conclude that studies that investigate the sources of person aberrance with real data should model the classroom as well as individual levels. [source] Multilevel Models in Family Research: Some Conceptual and Methodological IssuesJOURNAL OF MARRIAGE AND FAMILY, Issue 2 2002Jay Teachman Examining the impact of context on individual-level outcomes has become an increasingly common undertaking in the social sciences. The growth in concern for identifying the effects of macrolevel characteristics has generated both theoretical and methodological advancements. In this issue of Journal of Marriage and Family, Butler (2002) researches whether the effect of welfare benefit levels on premarital childbearing varies by context, Hoffmann (2002) researches the effect of context on adolescent drug use, and Simons et al. (2002) examine how the relationship between parenting and child conduct varies by context. These articles are used as a background to discuss important theoretical and methodological issues surrounding the analysis of multilevel data. The authors present a simple analysis of data pertaining to age at first marriage taken from the Panel Study of Income Dynamics and merged with census data to measure contextual effects as a pedagogical device for introducing readers to the benefits of multilevel modeling. [source] Multilevel models for estimating incremental net benefits in multinational studiesHEALTH ECONOMICS, Issue 8 2007Richard Grieve Abstract Multilevel models (MLMs) have been recommended for estimating incremental net benefits (INBs) in multicentre cost-effectiveness analysis (CEA). However, these models have assumed that the INBs are exchangeable and that there is a common variance across all centres. This paper examines the plausibility of these assumptions by comparing various MLMs for estimating the mean INB in a multinational CEA. The results showed that the MLMs that assumed the INBs were exchangeable and had a common variance led to incorrect inferences. The MLMs that included covariates to allow for systematic differences across the centres, and estimated different variances in each centre, made more plausible assumptions, fitted the data better and led to more appropriate inferences. We conclude that the validity of assumptions underlying MLMs used in CEA need to be critically evaluated before reliable conclusions can be drawn. Copyright © 2006 John Wiley & Sons, Ltd. [source] Longitudinal Associations Between Alcohol Problems and Depressive Symptoms: Early Adolescence Through Early AdulthoodALCOHOLISM, Issue 1 2009Naomi R. Marmorstein Background:, Alcohol use-related problems and depressive symptoms are clearly associated with each other, but results regarding the nature of this association have been inconsistent. In addition, the possible moderating effects of age and gender have not been comprehensively examined. The goals of this study were to clarify: (i) how depressive symptoms affect the levels and trajectory of alcohol use-related problems, (ii) how alcohol use-related problems affect the levels and trajectory of depressive symptoms, and (iii) whether there are differences in these associations at different points in development or between males and females. Methods:, Participants for this study were drawn from the National Longitudinal Study of Adolescent Health (AddHealth) data set, a community-based sample of 20,728 adolescents followed from adolescence through early adulthood. Multilevel models were used to assess how each problem affected the level and rate of change in the other problem over time; gender was considered as a possible moderator of these associations. Results:, The results indicated that alcohol use-related problems and depressive symptoms had reciprocal, positive associations with each other during the period from early adolescence through early adulthood; however, these effects differed somewhat by gender and age. High levels of depressive symptoms were associated with higher initial levels of alcohol problems (particularly among females), as well as faster increases in alcohol problems over time among males. High levels of alcohol problems were associated with higher initial levels of depressive symptoms (particularly among females), as well as less curvature in the slope of depressive symptoms so that although there was a large difference between people with high and low depressive symptoms in early adolescence, by early adulthood the difference was smaller (particularly among females). Conclusions:, These results highlight the importance of examining gender and age in studies on the associations between affective disorders and substance use disorders. [source] Multilevel models for longitudinal dataJOURNAL OF THE ROYAL STATISTICAL SOCIETY: SERIES A (STATISTICS IN SOCIETY), Issue 1 2008Fiona Steele Summary., Repeated measures and repeated events data have a hierarchical structure which can be analysed by using multilevel models. A growth curve model is an example of a multilevel random-coefficients model, whereas a discrete time event history model for recurrent events can be fitted as a multilevel logistic regression model. The paper describes extensions to the basic growth curve model to handle auto-correlated residuals, multiple-indicator latent variables and correlated growth processes, and event history models for correlated event processes. The multilevel approach to the analysis of repeated measures data is contrasted with structural equation modelling. The methods are illustrated in analyses of children's growth, changes in social and political attitudes, and the interrelationship between partnership transitions and childbearing. [source] Growth in root length of the mandibular canine and premolars in a mixed-longitudinal orthodontic sample,AMERICAN JOURNAL OF HUMAN BIOLOGY, Issue 5 2009Shelley L. Smith Numerous studies of dental development focusing on eruption (clinical emergence) exist in the literature, but fewer studies examine dental development as a process extending across years or decades, and root development is commonly assessed using fractional root lengths. Here, we examine the growth of mandibular canine and premolar roots in a mixed-longitudinal sample of orthodontic patients (77 females and 74 males) from north central Texas. Multilevel models are generated for root lengths as a percentage of total tooth lengths (within films) as well as for absolute root lengths (across films). As a percentage of tooth length, roots grow with decreasing velocity through time between 7 and 14 years of age. More complex patterns appear for absolute growth in root length, with girls showing an earlier maximum growth rate for the canine than for the premolars. Substitution of dental age for chronological age reduces between-subject variation (assessed at age 11), especially for boys. A better understanding of dental maturation, including root length growth, should allow improved prediction models. Am. J. Hum. Biol., 2009. © 2009 Wiley-Liss, Inc. [source] The Development of Gendered Interests and Personality Qualities From Middle Childhood Through Adolescence: A Biosocial AnalysisCHILD DEVELOPMENT, Issue 2 2009Susan M. McHale This study charted the development of gendered personality qualities and activity interests from age 7 to age 19 in 364 first- and secondborn siblings from 185 White, middle/working-class families, assessed links between time in gendered social contexts (with mother, father, female peers, and male peers) and gender development, and tested whether changes in testosterone moderated links between time use and gender development. Multilevel models documented that patterns of change varied across dimensions of gender and by sex and birth order and that time in gendered social contexts was generally linked to development of more stereotypical qualities. Associations between time with mother and expressivity and time with father and instrumentality were stronger for youth with slower increases in testosterone. [source] Peer Group Socialization of Homophobic Attitudes and Behavior During AdolescenceCHILD DEVELOPMENT, Issue 6 2007V. Paul Poteat A social developmental framework was applied to test for the socialization of homophobic attitudes and behavior within adolescent peer groups (Grades 7,11; aged 12,17 years). Substantial similarity within and differences across groups were documented. Multilevel models identified a group socializing contextual effect, predicting homophobic attitudes and behavior of individuals within the group 8 months later, even after controlling for the predictive effect of individuals' own previously reported attitudes and behavior. Several group characteristics moderated the extent to which individuals' previously reported attitudes predicted later attitudes. Findings indicate the need to integrate the concurrent assessment of individual and social factors to inform the construction of more comprehensive models of how prejudiced attitudes and behaviors develop and are perpetuated. [source] ESCAPING CRIME: THE EFFECTS OF DIRECT AND INDIRECT VICTIMIZATION ON MOVING,CRIMINOLOGY, Issue 4 2008MIN XIE This article investigates the impact of criminal victimization on household residential mobility. Existing research finds that direct experiences with crime influence mobility decisions, such that persons who suffer offenses near their homes are more likely to move. The current study extends this line of inquiry to consider whether indirect victimization that involves neighbors also stimulates moving. The analysis uses the National Crime Survey to estimate multilevel models that incorporate data from individual households and their spatially proximate neighbors. The results show that the link between direct victimization and moving continues to hold after controlling for neighborhood context. Indirect property victimization also leads to moving, with effects about equal in size to those of direct victimization. In contrast, no evidence is found that violent victimization that occurs in neighboring homes influences mobility, probably because most of these events are nonstranger violence that provokes less anxiety for neighbors. [source] Individuals receiving addiction treatment: are medical costs of their family members reduced?ADDICTION, Issue 7 2010Constance Weisner ABSTRACT Aims To examine whether alcohol and other drug (AOD) treatment is related to reduced medical costs of family members. Design Using the administrative databases of a private, integrated health plan, we matched AOD treatment patients with health plan members without AOD disorders on age, gender and utilization, identifying family members of each group. Setting Kaiser Permanente Northern California. Participants Family members of abstinent and non-abstinent AOD treatment patients and control family members. Measurements We measured abstinence at 1 year post-intake and examined health care costs per member-month of family members of AOD patients and of controls through 5 years. We used generalized estimating equation methods to examine differences in average medical cost per member-month for each year, between family members of abstinent and non-abstinent AOD patients and controls. We used multilevel models to examine 4-year cost trajectories, controlling for pre-intake cost, age, gender and family size. Results AOD patients' family members had significantly higher costs and more psychiatric and medical conditions than controls in the pre-treatment year. At 2,5 years, each year family members of AOD patients abstinent at 1 year had similar average per member-month medical costs to controls (e.g. difference at year 5 = $2.63; P > 0.82), whereas costs for family members of non-abstinent patients were higher (e.g. difference at year 5 = $35.59; P = 0.06). Family members of AOD patients not abstinent at 1 year, had a trajectory of increasing medical cost (slope = $10.32; P = 0.03) relative to controls. Conclusions Successful AOD treatment is related to medical cost reductions for family members, which may be considered a proxy for their improved health. [source] A structured and dynamic framework to advance traits-based theory and prediction in ecologyECOLOGY LETTERS, Issue 3 2010Colleen T. Webb Ecology Letters (2010) 13: 267,283 Abstract Predicting changes in community composition and ecosystem function in a rapidly changing world is a major research challenge in ecology. Traits-based approaches have elicited much recent interest, yet individual studies are not advancing a more general, predictive ecology. Significant progress will be facilitated by adopting a coherent theoretical framework comprised of three elements: an underlying trait distribution, a performance filter defining the fitness of traits in different environments, and a dynamic projection of the performance filter along some environmental gradient. This framework allows changes in the trait distribution and associated modifications to community composition or ecosystem function to be predicted across time or space. The structure and dynamics of the performance filter specify two key criteria by which we judge appropriate quantitative methods for testing traits-based hypotheses. Bayesian multilevel models, dynamical systems models and hybrid approaches meet both these criteria and have the potential to meaningfully advance traits-based ecology. [source] Using multilevel models for assessing the variability of multinational resource use and cost dataHEALTH ECONOMICS, Issue 2 2005Richard Grieve Abstract Multinational economic evaluations often calculate a single measure of cost-effectiveness using cost data pooled across several countries. To assess the validity of pooling international cost data the reasons for cost variation across countries need to be assessed. Previously, ordinary least-squares (OLS) regression models have been used to identify factors associated with variability in resource use and total costs. However, multilevel models (MLMs), which accommodate the hierarchical structure of the data, may be more appropriate. This paper compares these different techniques using a multinational dataset comprising case-mix, resource use and cost data on 1300 stroke admissions from 13 centres in 11 European countries. OLS and MLMs were used to estimate the effect of patient and centre-level covariates on the total length of hospital stay (LOS) and total cost. MLMs with normal and gamma distributions for the data within centres were compared. The results from the OLS model showed that both patient and centre-level covariates were associated with LOS and total cost. The estimates from the MLMs showed that none of the centre-level characteristics were associated with LOS, and the level of spending on health was the centre-level variable most highly associated with total cost. We conclude that using OLS models for assessing international variation can lead to incorrect inferences, and that MLMs are more appropriate for assessing why resource use and costs vary across centres. Copyright © 2004 John Wiley & Sons, Ltd. [source] Predicting river width, depth and velocity at ungauged sites in England and Wales using multilevel modelsHYDROLOGICAL PROCESSES, Issue 20 2008D. J. Booker Abstract Using a dataset of gauged river discharges taken from sites in England and Wales, linear multilevel models (also known as mixed effects models) were applied to quantify the variability in discharge and the discharge-hydraulic geometry relationships across three nested spatial scales. A jackknifing procedure was used to test the ability of the multilevel models to predict hydraulic geometry, and therefore width, mean depth and mean velocity, at ungauged stations. These models provide a framework for making predictions of hydraulic geometry parameters, with associated levels of uncertainty, using different levels of data availability. Results indicate that as one travels downstream along a river there is greater variability in hydraulic geometry than is the case between rivers of similar sizes. This indicates that hydraulic geometry (and therefore hydrology) is driven by catchment area, to a greater extent than by natural geomorphological variations in the streamwise direction at the mesoscale, but these geomorphological variations can still have a major impact on channel structure. Copyright © 2008 John Wiley & Sons, Ltd. [source] Multilevel investigation of variation in HoNOS ratings by mental health professionals: a naturalistic study of consecutive referralsINTERNATIONAL JOURNAL OF METHODS IN PSYCHIATRIC RESEARCH, Issue 3 2004R. Ecob Episodes of mental healthcare in specialist psychiatric services often begin with the assessment of clinical and psychosocial needs of patients by healthcare professionals. Particularly for patients with complex needs or severe problems, ratings of clinical and social functioning at the start of each episode of care may serve as a baseline against which subsequent measures can be compared. Currently, little is known about service variations in such assessments on referrals from primary care. We set out to quantify variability in initial assessments performed by healthcare professionals in three CMHTs in Bristol (UK) using the Health of the Nation Outcome Scales (HoNOS). We tested the hypothesis that variations in HoNOS total and sub-scale scores are related to referral source (general practices), healthcare assessor (in CMHTs) and the assessor's professional group. Statistical analysis was performed using multilevel variance components models with cross-classified random effects. We found that variation due to assessor substantially exceeded that due to referral source (general practices). Furthermore, patient variance differed by assessor profession for the HoNOS , Impairment scores. Assessor variance differed by assessor profession for the HoNOS , Social scores. As HoNOS total and subscale scores show much larger variation by assessor than by referral source, investigations of HoNOS scores must take assessors into account. Services should implement and evaluate interdisciplinary training to improve consistency in use of rating thresholds; such initiatives could be evaluated using these extensions of multilevel models. Future research should aim to integrate routine diagnostic data with continuous outcomes to address selection effects (of patients to assessors) better. Copyright © 2004 Whurr Publishers Ltd. [source] Religious Affiliation and Attendance Among Immigrants in Eight Western Countries: Individual and Contextual EffectsJOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION, Issue 1 2006FRANK VAN TUBERGEN This study examines the religious affiliation and participation of immigrants from a large-scale, comparative perspective. I propose a "specific migration" framework, in which immigrants' religiosity is an outcome of both individual characteristics and contextual properties related to immigrants' country of origin, country of destination, and combinations of origin and destination (i.e., communities). I use notions discussed in the religion and migration literature that fit into this scheme. To test these ideas, I collected and standardized 20 existing surveys on immigrants in eight Western countries, yielding about 38,000 immigrants. Applying multilevel models, I found, among other things, that: (1) immigrants from countries with higher levels of modernization express lower levels of religious commitment; (2) immigrants in religious countries are more religious themselves; and (3) the well-documented higher levels of religious commitment among women is not generalizable to immigrants. [source] Accounting for Aberrant Test Response Patterns Using Multilevel ModelsJOURNAL OF EDUCATIONAL MEASUREMENT, Issue 3 2007Alexandra Petridou Hypotheses about aberrant test-response behavior and hence invalid person-measurement have hitherto included factors like ability, gender, language, test-anxiety, and motivation, but these have not previously been collectively investigated with real data, or with multilevel models. This study analyzes the effect of these factors on person aberrance using a real mathematics assessment data set under the framework of a two-level (person and classroom) hierarchical model. The results suggest that higher-scoring pupils, and, to a lesser extent, second-language learners are significantly more often aberrant. But more importantly, we find that the classroom makes a significant contribution to person aberrance and conclude that studies that investigate the sources of person aberrance with real data should model the classroom as well as individual levels. [source] Longitudinal Dyad Models in Family ResearchJOURNAL OF MARRIAGE AND FAMILY, Issue 4 2005Karen S. Lyons Multilevel modeling allows for the simultaneous analysis of data gathered at more than 1 unit of analysis (e.g., children nested in schools). It is often used to examine the effects of various contexts on individual differences in change. This paper promotes the application of multilevel models to longitudinal dyadic data in family research. By focusing on the dyad as context, researchers can examine within-dyad change and begin to understand the interactive processes that constitute the relationship between partners. They can then frame questions about interdyad differences in within-dyad change. We present several longitudinal models that researchers can use to examine the pattern of change within dyads, assess heterogeneity in change across dyads, and investigate cross-partner effects on change. We comment on the implications of these models for family research. [source] The influence of work-family culture and workplace relationships on work interference with family: a multilevel modelJOURNAL OF ORGANIZATIONAL BEHAVIOR, Issue 7 2008Debra A. Major This research tested a multilevel model examining the influence of work-family culture and supportive workplace relationships on work interference with family. Web-based survey data were provided by 792 information technology employees from 10 organizations. Random coefficient modeling was used to test a path model examining the relationships between work-family culture, leader-member exchange (LMX), coworker support, and work interference with family. The direct effects of LMX and coworker support on work interference with family were significant. The indirect effect of work-family culture on work interference with family was also significant. Results demonstrate the value of work-family culture in understanding supportive supervisory and coworker relationships and work interference with family and highlight the need to employ multilevel models to understand these relationships. Copyright © 2007 John Wiley & Sons, Ltd. [source] The complexity of school and neighbourhood effects and movements of pupils on school differences in models of educational achievementJOURNAL OF THE ROYAL STATISTICAL SOCIETY: SERIES A (STATISTICS IN SOCIETY), Issue 3 2009George Leckie Summary., Traditional studies of school differences in educational achievement use multilevel modelling techniques to take into account the nesting of pupils within schools. However, educational data are known to have more complex non-hierarchical structures. The potential importance of such structures is apparent when considering the effect of pupil mobility during secondary schooling on educational achievement. Movements of pupils between schools suggest that we should model pupils as belonging to the series of schools that are attended and not just their final school. Since these school moves are strongly linked to residential moves, it is important to explore additionally whether achievement is also affected by the history of neighbourhoods that are lived in. Using the national pupil database, this paper combines multiple membership and cross-classified multilevel models to explore simultaneously the relationships between secondary school, primary school, neighbourhood and educational achievement. The results show a negative relationship between pupil mobility and achievement, the strength of which depends greatly on the nature and timing of these moves. Accounting for pupil mobility also reveals that schools and neighbourhoods are more important than shown by previous analysis. A strong primary school effect appears to last long after a child has left that phase of schooling. The additional effect of neighbourhoods, in contrast, is small. Crucially, the rank order of school effects across all types of pupil is sensitive to whether we account for the complexity of the multilevel data structure. [source] Multilevel models for longitudinal dataJOURNAL OF THE ROYAL STATISTICAL SOCIETY: SERIES A (STATISTICS IN SOCIETY), Issue 1 2008Fiona Steele Summary., Repeated measures and repeated events data have a hierarchical structure which can be analysed by using multilevel models. A growth curve model is an example of a multilevel random-coefficients model, whereas a discrete time event history model for recurrent events can be fitted as a multilevel logistic regression model. The paper describes extensions to the basic growth curve model to handle auto-correlated residuals, multiple-indicator latent variables and correlated growth processes, and event history models for correlated event processes. The multilevel approach to the analysis of repeated measures data is contrasted with structural equation modelling. The methods are illustrated in analyses of children's growth, changes in social and political attitudes, and the interrelationship between partnership transitions and childbearing. [source] Multilevel modelling of the number of property crimes: household and area effectsJOURNAL OF THE ROYAL STATISTICAL SOCIETY: SERIES A (STATISTICS IN SOCIETY), Issue 2 2006Andromachi Tseloni Summary., This study examines household and area effects on the incidence of total property crimes and burglaries and thefts. It uses data from the 2000 British Crime Survey and the 1991 UK census small area statistics. Results are obtained from estimated random-effects multilevel models, with an assumed negative binomial distribution of the dependent variable. Both household and area characteristics, as well as selected interactions, explain a significant portion of the variation in property crimes. There are also a large number of significant between-area random variances and covariances of household characteristics. The estimated fixed and random effects may assist in advancing victimization theory. The methods have potential for developing a better understanding of factors that give rise to crime and so assist in framing crime prevention policy. [source] Post-transplantation growth among pediatric recipients of liver transplantationPEDIATRIC TRANSPLANTATION, Issue 4 2005Idris V.R. Evans Abstract:, Improving a patient's quality-of-life (QOL) post-liver transplantation is of great importance. An aspect of improved QOL is the restoration of normal growth patterns in pediatric patients. To describe the post-transplantation growth patterns of 72 children included in the National Institute of Diabetes and Digestive and Kidney Diseases , Liver Transplantation Database (NIDDK-LTD), multilevel models were used, according to which children who waited more than a year for transplantation were smaller, compared with age and sex matched peers, at transplantation than children who waited less than a year while children who were growth retarded at transplantation experienced a larger yearly comparison height increase than children who were not growth retarded. The analysis also showed that boys older than 2 yr and younger than 13 yr at transplantation and girls older than 2 yr and younger than 11 yr at transplantation were significantly less growth retarded at transplantation than boys and girls under the age of 2 yr at transplantation. [source] Geographic variation in opioid prescribing for acute, work-related, low back pain and associated factors: A multilevel analysisAMERICAN JOURNAL OF INDUSTRIAL MEDICINE, Issue 2 2009Barbara S. Webster BSPT, PA-C Abstract Background Given reports about variation in opioid prescribing, concerns about increasing opioid use and its associated negative consequences make understanding the sources of variability important. The aims of the study were to assess the extent of and factors associated with geographic variation in early opioid prescribing for acute, work-related, low back pain (LBP). Methods Cases were selected from workers compensation administrative data filed between January 1, 2002 and December 31, 2003 and included claims from states with more than 40 cases. Early opioid prescribing (one or more prescriptions within first 15 days) was the outcome. Weighted coefficient of variation (wCOV) estimated geographic variation, and multilevel models measured variability controlling for individual and contextual factors. Results Of the 8,262 claimants, 21.3% received at least one early opioid prescription. Significant between-state variation was found (wCOV,=,53%), from 5.7% (Massachusetts) to 52.9% (South Carolina). Seventy-nine percent of the between-state variation was explained by three contextual factors: state household income inequality (prevalence ratio [PR] 1.06, 95% confidence interval [CI],=,1.01, 1.12), number of physicians per capita (PR 0.99, 95% CI,=,0.98, 0.99), and workers compensation cost containment effort score (PR 1.12, 95% CI,=,1.02, 1.24). Individual-level factors, including severity, explained only a small portion of the geographic variability. Conclusion Geographic variation of early opioid prescribing for acute LBP is important and almost fully explained by state-level contextual factors. The study suggests that clinician and patient interaction and the subsequent decision to use opioids are substantially framed by social conditions and control systems. Am. J. Ind. Med. 52:162,171, 2009. © 2008 Wiley-Liss, Inc. [source] Contextual Sources of AmbivalencePOLITICAL PSYCHOLOGY, Issue 5 2008Luke Keele When will people become ambivalent about politics? One possibility is that the roots of ambivalence lie within the individual, with differences in political knowledge and attitude strength predicting whether a person internalizes the conflicts of politics. Alternately, attitudinal ambivalence could result from structural differences in the way political choices are presented in the wider political environment. We explore the degree to which different environments promote or limit ambivalence using a matching approach in conjunction with a set of multilevel models. We find that campaign environments can induce candidate ambivalence. In presidential elections, campaign efforts promote ambivalence most when competition between partisan campaign efforts is high. In House elections, campaign spending has a direct effect on levels of candidate ambivalence, where a candidate's spending decreases ambivalence about that candidate and increases ambivalence about opponents. [source] Social stratification and attitudes: a comparative analysis of the effects of class and education in Europe1THE BRITISH JOURNAL OF SOCIOLOGY, Issue 4 2007Matthijs Kalmijn Abstract A classic topic in the sociology of inequality lies in the subjective consequences of people's stratification position. Many studies have shown that education and occupational class have significant effects on attitudes, but little is known about how the magnitude of these effects depends on the societal context. There has been debate in the scholarly literature, with some authors arguing that effects of class and education are less important when societies are more developed, whereas other authors argue that effects are either stable (for class) or increasing (for education). We use a meta-analytical design to address this debate. More specifically we examine the effects of class and education for a broad range of attitudes (21 scales) in 22 European countries using data from the 1999 wave of the European Values Study. We pool summary-measures of association (Eta-values) into a new dataset and analyse these Eta-values (N = 453) applying multilevel models with characteristics of countries and characteristics of attitudes as the independent variables. Our results show that there is no evidence that the effects of class on attitudes are lower when countries are more modern, but we do find larger effects of education in more modern countries. [source] When Should Epidemiologic Regressions Use Random Coefficients?BIOMETRICS, Issue 3 2000Sander Greenland Summary. Regression models with random coefficients arise naturally in both frequentist and Bayesian approaches to estimation problems. They are becoming widely available in standard computer packages under the headings of generalized linear mixed models, hierarchical models, and multilevel models. I here argue that such models offer a more scientifically defensible framework for epidemiologic analysis than the fixed-effects models now prevalent in epidemiology. The argument invokes an antiparsimony principle attributed to L. J. Savage, which is that models should be rich enough to reflect the complexity of the relations under study. It also invokes the countervailing principle that you cannot estimate anything if you try to estimate everything (often used to justify parsimony). Regression with random coefficients offers a rational compromise between these principles as well as an alternative to analyses based on standard variable-selection algorithms and their attendant distortion of uncertainty assessments. These points are illustrated with an analysis of data on diet, nutrition, and breast cancer. [source] Measuring Efficiency: A Comparison of Multilevel Modelling and Data Envelopment Analysis in the Context of Higher EducationBULLETIN OF ECONOMIC RESEARCH, Issue 2 2006Jill JohnesArticle first published online: 15 MAR 200 I21; C14; C16 Abstract Data envelopment analysis (DEA) and multilevel modelling (MLM) are applied to a data set of 54,564 graduates from UK universities in 1993 to assess whether the choice of technique affects the measurement of universities' performance. A methodology developed by Thanassoulis and Portela (2002; Education Economics, 10(2), pp. 183,207) allows each individual's DEA efficiency score to be decomposed into two components: one attributable to the university at which the student studied and the other attributable to the individual student. From the former component, a measure of each institution's teaching efficiency is derived and compared to the university effects from various multilevel models. The comparisons are made within four broad subjects: pure science, applied science, social science and arts. The results show that the rankings of universities derived from the DEA efficiencies which measure the universities' own performance (i.e., having excluded the efforts of the individuals) are not strongly correlated with the university rankings derived from the university effects of the multilevel models. The data were also used to perform a university-level DEA. The university efficiency scores derived from these DEAs are largely unrelated to the scores from the individual-level DEAs, confirming a result from a smaller data set (Johnes, 2006a; European Journal of Operational Research, forthcoming). However, the university-level DEAs provide efficiency scores which are generally strongly related to the university effects of the multilevel models. [source] | ||||||||||||||||||||||