In previous posts, I have describe in detail what constitutes Type I and Type II errors. This figure, however, conveys the concept much more succinctly. HT: Marginal Revolution.

Read the rest of this entry »## Increasing Generalizability of RCTs

Is your randomized controlled trial (RCT) generalizable to the general population? This question is known as external validity and is a major issue for a number of treatments. Sometimes, a treatment is very effective in an RCT, but less so in the real world. One reason why this may be the case is that the […]

Read the rest of this entry »## Measuring cause-specific mortality?

This question is not so easy to answer, even when using data from a randomized trial. Further, many studies do not have the statistical power to identify cause-specific mortality. Consider the following example from Kim and Thompson: Consider a trial of an intervention only influencing a single cause of death, or a few specific causes […]

Read the rest of this entry »## Beta Binomial Regression

Oftentimes, one will observe data cluster around two different points. This distribution is known as a bimodal distribution. A bimodal distribution could arise, for instance, when patients have two choices of health care providers, and the data measure the share of times patients use one of the providers. To model the effect of different covariates on variable […]

Read the rest of this entry »## Should we trust difference-in-difference estimators?

Difference-in-difference (DD) estimators are appropriate when the interventions are as good as random, conditional on time and group fixed effects. Although much of the debate in the economic literature regarding DD estimators focuses on the possible endogeneity of the interventions, there is another problem with DD: underestimated standard errors. A paper by Bertrand, Duflo and […]

Read the rest of this entry »## How to implement propensity score matching

What options are available for propensity score matching algorithms? Baser (2006) describes a number of popular options. Stratified Matching. In this method, the range of variation of the propensity score is divided into intervals such that within each interval, treated and control units have, on average, the same propensity score. Differences in outcome measures between […]

Read the rest of this entry »## Local Instrumental Variables

What is the effect of a treatment on health outcomes? The real question is: can you be more specific? Researchers may measure the treatment effect a variety of ways. Sensible research questions include: What is the average effect of the treatment across all individuals? What is the average treatment effect only among those who received […]

Read the rest of this entry »## Will missing data affect physician P4P scores?

The answer is yes, but maybe not as much as you may thought. A paper by Ryan and Bao use data from a randomized controlled trial (RCT) called IMPACT (Improving Mood-Promoting Access to Collaborative Treatment) to determine if errors in physician quality profiling are due mostly to random variation or missing data. For this report, […]

Read the rest of this entry »## Mahalanobis Distance

What is Mahalanobis distance? Most people know what Euclidean distance is…it is the shortest distance between any two points. In other words, its what we typically think of when we think of distance – the distance we would measure with a ruler, and the one given by the Pythagorean formula. Unlike Euclidean distance, Mahalanobis distance […]

Read the rest of this entry »