An approximate distribution of estimates of variance components.
By Satterthwaite Fe. 1946
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- Bernhard Schölkopf
- Jan Peters
- Peter Buhlmann
- Yoshua Bengio
- Kun Zhang
- Niklas Pfister
- Sergey Levine
- David Lopez-Paz
Citations in the corpus
- Kernel‐based tests for joint independence (2016, background): “on (see also Kankainen, 1995; Gretton et al., 2005) is that a distribution of the form P∞ i=1 λiZ 2 i can be approximated fairly well by a Gamma distribution with matched first and second moments (see Satterthwaite, 1946, for basic empirical evidence). This has, however, only been shown empirically a”
- Fitting Linear Mixed-Effects Models Using lme4 (2014, methodology): “Satterthwaite FE (1946).”
- HiCS: High Contrast Subspaces for Density-Based Outlier Ranking (2012, methodology): “, if both samples originate from the same probability density, the test statistic t follows a t-distribution with a degree of freedom which can be obtained by the Welch-Satterthwaite equation [27].”
- The power of statistical tests in meta-analysis. (2001, methodology): “When the conditional variances are unequal, an approximation to the distribution of Q can be derived using a method (Satterthwaite, 1946) that approximates the distribution of Q by a gamma random variable with mean and variance equal to that of Q.”
- A Practitioner’s Guide to Cluster-Robust Inference (2015, methodology): “Satterthwaite (1946) proposed an approximation that was extended to regression with clustered errors by Bell and McCaffrey (2002) and developed further by Imbens and Kolesar (2012).”
- A Robust Test for Weak Instruments (2013, methodology): “Empirical researchers frequently report the robust F statistic as a simple way of adjusting the Staiger and Stock (1997) and Stock and Yogo (2005) pre-tests for heteroskedasticity, serial correlation, and clustering, and compare them to the homoskedastic critical values.”
- The power of statistical tests for moderators in meta-analysis. (2004, methodology): “05 level test that the residual variance component is 0 using Satterthwaite’s (1946) approximation.”