PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 9285235

Cluster effects and simultaneity in multilevel models.

R Blundell·F Windmeijer
View original source ↗Export metadata ↓

Abstract

For small group sizes, the GLS estimator in multilevel models is biased and inconsistent when the random cluster effects are correlated with the regressors. A fixed effects approach, conditioning on the cluster effects, provides consistent estimates for the slope parameters. The two estimators are equivalent when group sizes are large. The same results obtain for two-stage estimation procedures that allow for some of the regressors to be simultaneously determined with the dependent variable. The GLS and fixed effects estimators are applied to data on acute care hospital utilization in the UK, allowing for health authority district effects.

Explore related subjects

Effect Modifier, EpidemiologicHealth Services ResearchHumansLeast-Squares AnalysisSample SizeJournal ArticleResearch Support, Non-U.S. Gov't

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

R Blundell, F Windmeijer. Cluster effects and simultaneity in multilevel models.. https://doi.org/10.1002/(sici)1099-1050(199707)6%3A4%3C439%3A%3Aaid-hec276%3E3.0.co%3B2-b

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed1997

Effect size and power for clinical trials that measure years of healthy life.

Some clinical trials perform repeated measurements on patients over time, plot those measures against time, and summarize the results in terms of the area under the curve. If the measured variable is health status, the summary outcome is sometimes referred to as years of healthy life (YHL), or quality-adjusted life years (QALY). This paper investigates some theoretical and practical aspects of randomized trials designed to assess measures such as YHL. We first derived algebraic expressions for the effect size of YHL measures under several theoretical models of the treatment's effect on health. We used these expressions to examine how the length of the study, the number of measurements per person and the correlations among health measurements over time influence the effect size. We also explored the relative statistical power of analyses based on YHL versus analyses based on change-scores using the same data. We present an example. Findings suggest that: (i) the number of measurements per person need not be large; (ii) high correlation among measures over time tends to lower the power of a study using YHL; (iii) a longer study will not always provide more power than a shorter study, and (iv) analyses based on YHL may have less power than change-score analyses. Some of these findings depend on the model of change in health status caused by the treatment. Such models require further study.

Effect Modifier, Epidemiologic↗
PubMed1997

Incorporating variability in estimates of heterogeneity in the random effects model in meta-analysis.

When combining results from separate investigations in a meta-analysis, random effects methods enable the modelling of differences between studies by incorporating a heterogeneity parameter tau 2 that accounts explicitly for across-study variation. We develop a simple form for the variance of Cochran's homogeneity statistic Q, leading to interval estimation of tau 2 utilizing an approximating distribution for Q; this enables us to extend the point estimation of DerSimonian and Laird. We also develop asymptotic likelihood methods and compared them with this method. We then use these approximating distributions to give a new method of calculating the weight given to the individual studies' results when estimating the overall mean which takes into account variation in these point estimates of tau 2. Two examples illustrate the methods presented, where we show that the new weighting scheme is between the standard fixed and random effects models in down-weighting the results of large studies and up-weighting those of small studies.

Effect Modifier, Epidemiologic↗