The Balance-Sample Size Frontier in Matching Methods for Causal Inference
We propose a simplified approach to matching for causal inference that simultaneously optimizes balance (similarity between the treated and control groups) and matched sample size. Existing approaches either fix the matched sample size and maximize balance or fix balance and maximize sample size, le...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
Wiley
2018
|
| Online Access: | http://hdl.handle.net/1721.1/118640 https://orcid.org/0000-0003-0205-5227 |
| _version_ | 1826191904792903680 |
|---|---|
| author | King, Gary Lucas, Christopher Nielsen, Richard Alexander |
| author2 | Massachusetts Institute of Technology. Department of Political Science |
| author_facet | Massachusetts Institute of Technology. Department of Political Science King, Gary Lucas, Christopher Nielsen, Richard Alexander |
| author_sort | King, Gary |
| collection | MIT |
| description | We propose a simplified approach to matching for causal inference that simultaneously optimizes balance (similarity between the treated and control groups) and matched sample size. Existing approaches either fix the matched sample size and maximize balance or fix balance and maximize sample size, leaving analysts to settle for suboptimal solutions or attempt manual optimization by iteratively tweaking their matching method and rechecking balance. To jointly maximize balance and sample size, we introduce the matching frontier, the set of matching solutions with maximum possible balance for each sample size. Rather than iterating, researchers can choose matching solutions from the frontier for analysis in one step. We derive fast algorithms that calculate the matching frontier for several commonly used balance metrics. We demonstrate this approach with analyses of the effect of sex on judging and job training programs that show how the methods we introduce can extract new knowledge from existing data sets. |
| first_indexed | 2024-09-23T09:03:07Z |
| format | Article |
| id | mit-1721.1/118640 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T09:03:07Z |
| publishDate | 2018 |
| publisher | Wiley |
| record_format | dspace |
| spelling | mit-1721.1/1186402022-09-30T13:07:42Z The Balance-Sample Size Frontier in Matching Methods for Causal Inference King, Gary Lucas, Christopher Nielsen, Richard Alexander Massachusetts Institute of Technology. Department of Political Science Nielsen, Richard Alexander We propose a simplified approach to matching for causal inference that simultaneously optimizes balance (similarity between the treated and control groups) and matched sample size. Existing approaches either fix the matched sample size and maximize balance or fix balance and maximize sample size, leaving analysts to settle for suboptimal solutions or attempt manual optimization by iteratively tweaking their matching method and rechecking balance. To jointly maximize balance and sample size, we introduce the matching frontier, the set of matching solutions with maximum possible balance for each sample size. Rather than iterating, researchers can choose matching solutions from the frontier for analysis in one step. We derive fast algorithms that calculate the matching frontier for several commonly used balance metrics. We demonstrate this approach with analyses of the effect of sex on judging and job training programs that show how the methods we introduce can extract new knowledge from existing data sets. 2018-10-22T15:00:47Z 2018-10-22T15:00:47Z 2016-11 2018-10-17T18:19:03Z Article http://purl.org/eprint/type/JournalArticle 00925853 http://hdl.handle.net/1721.1/118640 King, Gary, Christopher Lucas, and Richard A. Nielsen. “The Balance-Sample Size Frontier in Matching Methods for Causal Inference.” American Journal of Political Science 61, no. 2 (November 9, 2016): 473–489. https://orcid.org/0000-0003-0205-5227 http://dx.doi.org/10.1111/AJPS.12272 American Journal of Political Science Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Wiley Other repository |
| spellingShingle | King, Gary Lucas, Christopher Nielsen, Richard Alexander The Balance-Sample Size Frontier in Matching Methods for Causal Inference |
| title | The Balance-Sample Size Frontier in Matching Methods for Causal Inference |
| title_full | The Balance-Sample Size Frontier in Matching Methods for Causal Inference |
| title_fullStr | The Balance-Sample Size Frontier in Matching Methods for Causal Inference |
| title_full_unstemmed | The Balance-Sample Size Frontier in Matching Methods for Causal Inference |
| title_short | The Balance-Sample Size Frontier in Matching Methods for Causal Inference |
| title_sort | balance sample size frontier in matching methods for causal inference |
| url | http://hdl.handle.net/1721.1/118640 https://orcid.org/0000-0003-0205-5227 |
| work_keys_str_mv | AT kinggary thebalancesamplesizefrontierinmatchingmethodsforcausalinference AT lucaschristopher thebalancesamplesizefrontierinmatchingmethodsforcausalinference AT nielsenrichardalexander thebalancesamplesizefrontierinmatchingmethodsforcausalinference AT kinggary balancesamplesizefrontierinmatchingmethodsforcausalinference AT lucaschristopher balancesamplesizefrontierinmatchingmethodsforcausalinference AT nielsenrichardalexander balancesamplesizefrontierinmatchingmethodsforcausalinference |