I mostly write about statistics, computing, and academia. Subscribe to my atom feed!

A performance increase for partial dependence in edarf

Peer-Review is a Weak Signal of Paper Quality, Aug 2016peer-review is a high variance estimator of paper quality

Academic Technical Debt, Aug 2016academic code has technical debt

Generalization Error for Model Validation, Jul 2016generalization error should be used for model evaluation more frequently

Salsa Vaya Review, Jul 2016me talking about my bike

set up Emacs, R, Python, and Julia with Homebrew

An Analysis of Polity IV and its Components, Oct 2013Gleditsch and Ward (1997) aren't quite correct and Polity is mis/over-used

GNU Make for Reproducible Data Analysis, Nov 2013how to use GNU Make to make your research more reproducible

Analyzing Data from the Titanic, Jan 2013predicting who survived the Titanic using a GLM and a GAM

Statistical Education in Political Science, May 2014research design, probability, and mathematical statistics are necessary to a good statistical education

generalization error should be used for model evaluation more frequently

Monte Carlo for Teaching Applied Statistics, Dec 2014real data is overused in applied statistics classes in political science

Data Contingent Analysis of Historical Data, Oct 2014analysis of historical data is contingent on the data (duh!)

On the Epistemology of Statistics, Sep 2014articles that changed my thinking on what/how we can learn from data

Real-World Examples in Methods Papers, Jul 2014real-world examples in methods papers should be examples for applied researchers to follow

Statistical Education in Political Science, May 2014research design, probability, and mathematical statistics are necessary to a good statistical education

Data Mining as Exploratory Data Analysis, Feb 2014statistical learning should often be used in observational studies w/o (as-if) randomization

Identification, Public Policy, and Marijuana Legalization, Dec 2013causal inference (necessary for policy recommendations imo) with observational data is often hard or impossible

Analyzing Data from the Titanic, Jan 2013predicting who survived the Titanic using a GLM and a GAM

peer-review is a high variance estimator of paper quality

On Academic Publishers, Dec 2013publishing could be nearly free and for-profit publishers don't add value to academic research

Stress Physiology and Graduate School, Jan 2012a summary of "Why Zebra's Don't Get Ulcers" (a book on stress) adapted for graduate students

why doesn't the Powell Doctrine work? (it's politicians)

me talking about my bike

real data is overused in applied statistics classes in political science

Statistical Education in Political Science, May 2014research design, probability, and mathematical statistics are necessary to a good statistical education

Math Education and Political Science, Nov 2012math camp is no substitute for years of undergraduate math

A performance increase for partial dependence in edarf

Academic Technical Debt, Aug 2016academic code has technical debt

Git and GitHub Tutorial, Jan 2015the basics of how to set up and use git (from the cli) and github for scientists

GNU Make for Reproducible Data Analysis, Nov 2013how to use GNU Make to make your research more reproducible

Setting up OSX for Data Analysis, Jun 2013set up Emacs, R, Python, and Julia with Homebrew

Git/GitHub, Transparency, and Legitimacy in Quantitative Research, Mar 2013revision control (git) can help make research more transparent and reproducible

Analyzing Data from the Titanic, Jan 2013predicting who survived the Titanic using a GLM and a GAM

safari and hiking/climbing in Kenya

Washington, May 2015climbing in the Cascades

Ecuador, Jan 2015hiking and mountaineering in Ecuador

Boulder, Colorado, Jun 2014hiking and climbing in and around Boulder

A performance increase for partial dependence in edarf

Peer-Review is a Weak Signal of Paper Quality, Aug 2016peer-review is a high variance estimator of paper quality

Academic Technical Debt, Aug 2016academic code has technical debt

Generalization Error for Model Validation, Jul 2016generalization error should be used for model evaluation more frequently

Redux of An Empirical Evaluation of Explanations for State Repression, Oct 2014analysis of additional (relative to eeesr) predictors of state repression

An Empirical Evaluation of Explanations for State Repression, Apr 2014null-hypothesis significance testing is over/mis-used, predictive validity is important, and democracy/civil conflict partially measure state repression

Data Mining as Exploratory Data Analysis, Feb 2014statistical learning should often be used in observational studies w/o (as-if) randomization

Identification, Public Policy, and Marijuana Legalization, Dec 2013causal inference (necessary for policy recommendations imo) with observational data is often hard or impossible

An Analysis of Polity IV and its Components, Oct 2013Gleditsch and Ward (1997) aren't quite correct and Polity is mis/over-used