![](https://d3ilqtpdwi981i.cloudfront.net/GH0qFRex45FST3du_5BfI-m8_xc=/425x550/smart/https://bepress-attached-resources.s3.amazonaws.com/uploads/3f/70/9e/3f709ea8-005d-4f9a-a3ee-a94c2bcaf345/thumbnail_01284649-2d06-4519-b947-bae67b414eb9.jpg)
Many studies include some form of blocking in the study design. Block effects are rarely of intrinsic interest; instead they are included in a model so that that model reflects the study design. I consider the question of how these block effects should be modeled: as fixed effects or as random effects. I discuss the consequences of the choice, including the recovery of inter-block information when available, give a simple example to illustrate the connection between recovery of inter-block information and pooling two estimators of a treatment effect, and give an example where fitting a model with random block effects can lead to the wrong answer. I suggest that block effects should be modeled as fixed effects unless there are compelling reasons to do otherwise.
Available at: http://works.bepress.com/philip-dixon/64/
This proceeding is from Dixon, P.M. Should blocks be fixed or random? 2016 Conference on Applied Statistics in Agriculture Proceedings, May 1-3, Manhattan Kansas. Kansas State University. doi: 10.4148/2475-7772.1474.