scholarly journals The effect of number of clusters and cluster size on statistical power and Type I error rates when testing random effects variance components in multilevel linear and logistic regression models

2018 ◽  
Vol 88 (16) ◽  
pp. 3151-3163 ◽  
Author(s):  
Peter C. Austin ◽  
George Leckie
2019 ◽  
Author(s):  
Joshua Nugent ◽  
Ken Kleinman

Abstract Background: Linear mixed models (LMM) are a common approach to analyzing data from cluster randomized trials (CRTs). Inference on parameters can be performed via Wald tests or likelihood ratio tests (LRT), but both approaches may give incorrect Type I error rates in common finite sample settings. The impact of interactions of cluster size, number of clusters, intraclass correlation coefficient (ICC), and analysis approach on Type I error rates have not been well studied. Reviews of published CRTs find that small sample sizes are not uncommon, so the performance of different inferential approaches in these settings can guide data analysts to the best choices. Methods: Using a random-intercept LMM stucture, we use simulations to study Type I error rates with the LRT and Wald test with different degrees of freedom (DF) choices across different combinations of cluster size, number of clusters, and ICC.Results: Our simulations show that the LRT can be anti-conservative when the ICC is large and the number of clusters is small, with the effect most pronouced when the cluster size is relatively large. Wald tests with the Between-Within DF method or the Satterthwaite DF approximation maintain Type I error control at the stated level, though they are conservative when the number of clusters, the cluster size, and the ICC are small. Conclusions: Depending on the structure of the CRT, analysts should choose a hypothesis testing approach that will maintain the appropriate Type I error rate for their data. Wald tests with the Satterthwaite DF approximation work well in many circumstances, but in other cases the LRT may have Type I error rates closer to the nominal level.


2020 ◽  
Author(s):  
Joshua Nugent ◽  
Ken Kleinman

Abstract Background: Linear mixed models (LMM) are a common approach to analyzing data from cluster randomized trials (CRTs). Inference on parameters can be performed via Wald tests or likelihood ratio tests (LRT), but both approaches may give incorrect Type I error rates in common finite sample settings. The impact of different combinations of cluster size, number of clusters, intraclass correlation coefficient (ICC), and analysis approach on Type I error rates has not been well studied. Reviews of published CRTs nd that small sample sizes are not uncommon, so the performance of different inferential approaches in these settings can guide data analysts to the best choices.Methods: Using a random-intercept LMM stucture, we use simulations to study Type I error rates with the LRT and Wald test with different degrees of freedom (DF) choices across different combinations of cluster size, number of clusters, and ICC.Results: Our simulations show that the LRT can be anti-conservative when the ICC is large and the number of clusters is small, with the effect most pronounced when the cluster size is relatively large. Wald tests with the between-within DF method or the Satterthwaite DF approximation maintain Type I error control at the stated level, though they are conservative when the number of clusters, the cluster size, and the ICC are small.Conclusions: Depending on the structure of the CRT, analysts should choose a hypothesis testing approach that will maintain the appropriate Type I error rate for their data. Wald tests with the Satterthwaite DF approximation work well in many circumstances, but in other cases the LRT may have Type I error rates closer to the nominal level.


2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Joshua R. Nugent ◽  
Ken P. Kleinman

Abstract Background Linear mixed models (LMM) are a common approach to analyzing data from cluster randomized trials (CRTs). Inference on parameters can be performed via Wald tests or likelihood ratio tests (LRT), but both approaches may give incorrect Type I error rates in common finite sample settings. The impact of different combinations of cluster size, number of clusters, intraclass correlation coefficient (ICC), and analysis approach on Type I error rates has not been well studied. Reviews of published CRTs find that small sample sizes are not uncommon, so the performance of different inferential approaches in these settings can guide data analysts to the best choices. Methods Using a random-intercept LMM stucture, we use simulations to study Type I error rates with the LRT and Wald test with different degrees of freedom (DF) choices across different combinations of cluster size, number of clusters, and ICC. Results Our simulations show that the LRT can be anti-conservative when the ICC is large and the number of clusters is small, with the effect most pronouced when the cluster size is relatively large. Wald tests with the between-within DF method or the Satterthwaite DF approximation maintain Type I error control at the stated level, though they are conservative when the number of clusters, the cluster size, and the ICC are small. Conclusions Depending on the structure of the CRT, analysts should choose a hypothesis testing approach that will maintain the appropriate Type I error rate for their data. Wald tests with the Satterthwaite DF approximation work well in many circumstances, but in other cases the LRT may have Type I error rates closer to the nominal level.


2018 ◽  
Vol 20 (6) ◽  
pp. 2055-2065 ◽  
Author(s):  
Johannes Brägelmann ◽  
Justo Lorenzo Bermejo

Abstract Technological advances and reduced costs of high-density methylation arrays have led to an increasing number of association studies on the possible relationship between human disease and epigenetic variability. DNA samples from peripheral blood or other tissue types are analyzed in epigenome-wide association studies (EWAS) to detect methylation differences related to a particular phenotype. Since information on the cell-type composition of the sample is generally not available and methylation profiles are cell-type specific, statistical methods have been developed for adjustment of cell-type heterogeneity in EWAS. In this study we systematically compared five popular adjustment methods: the factored spectrally transformed linear mixed model (FaST-LMM-EWASher), the sparse principal component analysis algorithm ReFACTor, surrogate variable analysis (SVA), independent SVA (ISVA) and an optimized version of SVA (SmartSVA). We used real data and applied a multilayered simulation framework to assess the type I error rate, the statistical power and the quality of estimated methylation differences according to major study characteristics. While all five adjustment methods improved false-positive rates compared with unadjusted analyses, FaST-LMM-EWASher resulted in the lowest type I error rate at the expense of low statistical power. SVA efficiently corrected for cell-type heterogeneity in EWAS up to 200 cases and 200 controls, but did not control type I error rates in larger studies. Results based on real data sets confirmed simulation findings with the strongest control of type I error rates by FaST-LMM-EWASher and SmartSVA. Overall, ReFACTor, ISVA and SmartSVA showed the best comparable statistical power, quality of estimated methylation differences and runtime.


2017 ◽  
Vol 78 (3) ◽  
pp. 460-481 ◽  
Author(s):  
Margarita Olivera-Aguilar ◽  
Samuel H. Rikoon ◽  
Oscar Gonzalez ◽  
Yasemin Kisbu-Sakarya ◽  
David P. MacKinnon

When testing a statistical mediation model, it is assumed that factorial measurement invariance holds for the mediating construct across levels of the independent variable X. The consequences of failing to address the violations of measurement invariance in mediation models are largely unknown. The purpose of the present study was to systematically examine the impact of mediator noninvariance on the Type I error rates, statistical power, and relative bias in parameter estimates of the mediated effect in the single mediator model. The results of a large simulation study indicated that, in general, the mediated effect was robust to violations of invariance in loadings. In contrast, most conditions with violations of intercept invariance exhibited severely positively biased mediated effects, Type I error rates above acceptable levels, and statistical power larger than in the invariant conditions. The implications of these results are discussed and recommendations are offered.


2017 ◽  
Vol 88 (4) ◽  
pp. 769-784
Author(s):  
Falynn C. Turley ◽  
David Redden ◽  
Janice L. Case ◽  
Charles Katholi ◽  
Jeff Szychowski ◽  
...  

2016 ◽  
Author(s):  
Daijiang Li ◽  
Anthony R Ives

1. A growing number of studies incorporate functional trait information to analyse patterns and processes of community assembly. These studies of trait-environment relationships generally ignore phylogenetic relationships among species. When functional traits and the residual variation in species distributions among communities have phylogenetic signal, however, analyses ignoring phylogenetic relationships can decrease estimation accuracy and power, inflate type I error rates, and lead to potentially false conclusions. 2. Using simulations, we compared estimation accuracy, statistical power, and type I error rates of linear mixed models (LMM) and phylogenetic linear mixed models (PLMM) designed to test for trait-environment interactions in the distribution of species abundances among sites. We considered the consequences of both phylogenetic signal in traits and phylogenetic signal in the residual variation of species distributions generated by an unmeasured (latent) trait with phylogenetic signal. 3. When there was phylogenetic signal in the residual variation of species among sites, PLMM provided better estimates (closer to the true value) and greater statistical power for testing whether the trait-environment interaction regression coefficient differed from zero. LMM had unacceptably high type I error rates when there was phylogenetic signal in both traits and the residual variation in species distributions. When there was no phylogenetic signal in the residual variation in species distributions, LMM and PLMM had similar performances. 4. LMMs that ignore phylogenetic relationships can lead to poor statistical tests of trait-environment relationships when there is phylogenetic signal in the residual variation of species distributions among sites, such as caused by unmeasured traits. Therefore, phylogenies and PLMMs should be used when studying how functional traits affect species abundances among communities in response to environmental gradients.


Sign in / Sign up

Export Citation Format

Share Document