Teaching hypothesis testing with simulated distributions

Incorporating Racial Justice Topics into an Econometrics Course

“… 5 These papers include Briand and Hill ( 2013 ) on using Monte Carlo experiments to teach about the least squares estimator unbiasedness and confidence intervals, O’Hara ( 2014 ) on implementing instructor-written macros to teach sampling using bootstrapping examples, Hansen ( 2017 ) on teaching time-series econometrics, O’Hara ( 2019 ) on utilizing simulated distributions to teach hypothesis testing, and Stewart ( 2019 ) on teaching simultaneous equations. …”

Consilience of methods for phylogenetic analysis of variance

“…Performed many times, either approach allows a sampling distribution proxy (of a real but perhaps intractable sampling distribution) of a test statistic to be empirically generated, and inferences about the observed data are made based on the location of the observed statistic in the sampling distribution. It has been shown that empirical sampling distributions obtained from simulation‐based inference approaches can accurately approximate both likelihood profiles (Diggle and Gratton 1984; Gourieroux and Monfort 1993) and theoretical sampling distributions of summary statistics (Kac 1949; O'Hara 2019) for models that could equally use parametric probability distributions as proxies for sampling distributions. Likewise, various tests based on resampling data or residuals of linear models have been shown to have good statistical properties in terms of type I error, statistical power, and asymptotic convergence on exact tests (Anderson and Robinson 2001; Manly 2007).…”