Rob Olsen

Research Professor at George Washington University. Independent consultant on impact evaluation design, analysis, & methods. Expert on generalizability.

[Rob Olsen @EvalRob]

Looking for research on the factors that influence whether people or sites agree to participate in research studies when asked to do so. And even better, research on effective strategies for gaining cooperation. If you have any leads, please share! And thanks in advance.

[Rob Olsen @EvalRob]

Need to select a two-stage probability sample of districts and schools for an education study (or any two-stage sample)? Check out sitepickR, our new R package. Thanks to my collaborators, @Lizstuartdc and Elena Badillo-Goicoechea! https://t.co/lRSEExcVP8\

[Rob Olsen @EvalRob]

@Lizstuartdc @stats_tipton To Liz's point, the reason it would be useful to name districts and schools in RCTs is to help readers assess the generalizability of the RCT findings. Curious to better understand possible IRB objections.

[Rob Olsen @EvalRob]

Why do RCTs in education rarely identify the districts or schools that participate? Research by @Lizstuartdc and @stats_tipton finds that these districts and schools have distinctive characteristics. But it’s hard to identify them from journal articles and reports—any idea why?