Properties of restricted randomization with implications for experimental design
Recently, there has been increasing interest in the use of heavily restricted randomization designs which enforce balance on observed covariates in randomized controlled trials. However, when restrictions are strict, there is a risk that the treatment effect estimator will have a very high mean squa...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
De Gruyter
2022-09-01
|
| Series: | Journal of Causal Inference |
| Subjects: | |
| Online Access: | https://doi.org/10.1515/jci-2021-0057 |
| _version_ | 1818027426597306368 |
|---|---|
| author | Nordin Mattias Schultzberg Mårten |
| author_facet | Nordin Mattias Schultzberg Mårten |
| author_sort | Nordin Mattias |
| collection | DOAJ |
| description | Recently, there has been increasing interest in the use of heavily restricted randomization designs which enforce balance on observed covariates in randomized controlled trials. However, when restrictions are strict, there is a risk that the treatment effect estimator will have a very high mean squared error (MSE). In this article, we formalize this risk and propose a novel combinatoric-based approach to describe and address this issue. First, we validate our new approach by re-proving some known properties of complete randomization and restricted randomization. Second, we propose a novel diagnostic measure for restricted designs that only use the information embedded in the combinatorics of the design. Third, we show that the variance of the MSE of the difference-in-means estimator in a randomized experiment is a linear function of this diagnostic measure. Finally, we identify situations in which restricted designs can lead to an increased risk of getting a high MSE and discuss how our diagnostic measure can be used to detect such designs. Our results have implications for any restricted randomization design and can be used to evaluate the trade-off between enforcing balance on observed covariates and avoiding too restrictive designs. |
| first_indexed | 2024-12-10T04:47:43Z |
| format | Article |
| id | doaj.art-4b2dada8450442feb3ef69690c098c18 |
| institution | Directory Open Access Journal |
| issn | 2193-3685 |
| language | English |
| last_indexed | 2024-12-10T04:47:43Z |
| publishDate | 2022-09-01 |
| publisher | De Gruyter |
| record_format | Article |
| series | Journal of Causal Inference |
| spelling | doaj.art-4b2dada8450442feb3ef69690c098c182022-12-22T02:01:42ZengDe GruyterJournal of Causal Inference2193-36852022-09-0110122724510.1515/jci-2021-0057Properties of restricted randomization with implications for experimental designNordin Mattias0Schultzberg Mårten1Department of Statistics, Uppsala Center for Fiscal Studies (UCFS) and Urban Lab, Uppsala University, Uppsala, SwedenSpotify and Department of Statistics, Uppsala University, Uppsala, SwedenRecently, there has been increasing interest in the use of heavily restricted randomization designs which enforce balance on observed covariates in randomized controlled trials. However, when restrictions are strict, there is a risk that the treatment effect estimator will have a very high mean squared error (MSE). In this article, we formalize this risk and propose a novel combinatoric-based approach to describe and address this issue. First, we validate our new approach by re-proving some known properties of complete randomization and restricted randomization. Second, we propose a novel diagnostic measure for restricted designs that only use the information embedded in the combinatorics of the design. Third, we show that the variance of the MSE of the difference-in-means estimator in a randomized experiment is a linear function of this diagnostic measure. Finally, we identify situations in which restricted designs can lead to an increased risk of getting a high MSE and discuss how our diagnostic measure can be used to detect such designs. Our results have implications for any restricted randomization design and can be used to evaluate the trade-off between enforcing balance on observed covariates and avoiding too restrictive designs.https://doi.org/10.1515/jci-2021-0057experimental designrestricted randomizationrerandomizationcomputationally intensive methods62c2062k1062k99 |
| spellingShingle | Nordin Mattias Schultzberg Mårten Properties of restricted randomization with implications for experimental design Journal of Causal Inference experimental design restricted randomization rerandomization computationally intensive methods 62c20 62k10 62k99 |
| title | Properties of restricted randomization with implications for experimental design |
| title_full | Properties of restricted randomization with implications for experimental design |
| title_fullStr | Properties of restricted randomization with implications for experimental design |
| title_full_unstemmed | Properties of restricted randomization with implications for experimental design |
| title_short | Properties of restricted randomization with implications for experimental design |
| title_sort | properties of restricted randomization with implications for experimental design |
| topic | experimental design restricted randomization rerandomization computationally intensive methods 62c20 62k10 62k99 |
| url | https://doi.org/10.1515/jci-2021-0057 |
| work_keys_str_mv | AT nordinmattias propertiesofrestrictedrandomizationwithimplicationsforexperimentaldesign AT schultzbergmarten propertiesofrestrictedrandomizationwithimplicationsforexperimentaldesign |