The problem of false positives: Antibody tests in times of Corona

A few weeks or months from now, we could have a Covid-19 test kit sent to our home. Similar to a blood sugar test for diabetics, we would prick our finger, wait for a couple of minutes, and we will know whether we are immune or not. The general idea is that this would help in lessen the social distancing restrictions because those who are immune could in principle go back to a normal life.… Read More »The problem of false positives: Antibody tests in times of Corona

Estimating power for structural equation models: Simulations in R

Computing a priori power analyses for simple statistical models can be done analytically (e.g., with G*Power or the pwr package in R). However, estimating the power for more complex models and in particular structural equation models (SEM) is not as straightforward and requires simulations. I recently came across the package paramtest (Hughes, 2017) which provides a great framework for conducting more complex power simulations. In what follows, I provide some examples of how to simulate… Read More »Estimating power for structural equation models: Simulations in R

How to do specification curve analyses in R: Introducing ‘specr’

In the last month, Michael Scharkow and I have worked on a new R-package called specr. The goal was to facilitate specification curve analyses (also called multiverse analyses). The idea behind a specification curve analysis stems from the observation that a researcher has many degrees of freedom when conducting a quantitative analysis of a data set and sometimes, we do not really know how different decisions may impact the results. It starts with the question… Read More »How to do specification curve analyses in R: Introducing ‘specr’

How to visualize interaction effects

I recently gave a workshop on data visualization. One of the topics was visualizing interaction effects (or “moderation” analyses). I think it is a great topic because it exemplifies quite well that there is not one solution to all problems. In general, I would argue that trying to visualize interaction effects is great idea. Due to the conditional nature of the effects obtained from standard regression analyses that include an interaction term, it is often… Read More »How to visualize interaction effects

How to center in multilevel models

Have you ever thought about centering your variables before running an regression based analysis? From my personal experiences, chances are high are that you haven’t: Although a useful data transformation procedure for many statistical analyses, centering is seldom taught in fundamental statistic courses. Although the mathematical principles behind it may seem arbitrary, its implications are oftentimes quite strong. Although centering is already useful in standard statistical modeling (e.g., OLS regression), its usefulness is particularly evident… Read More »How to center in multilevel models