(test) doi:10.7910/DVN/26935

Latest Version
{
    "title": "Replication data for: How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It", 
    "id": 53121, 
    "persistentIdInfo": {
        "protocol": "doi", 
        "authority": "10.7910/DVN", 
        "identifier": "26935", 
        "persistentId": "doi:10.7910/DVN/26935", 
        "persistentUrl": "http://dx.doi.org/10.7910/DVN/26935"
    }, 
    "isPublished": true, 
    "timestamps": {
        "createdate": "2014-08-04 12:44:02", 
        "publicationdate": "2014-08-04 12:52:14"
    }, 
    "semanticVersionInfo": {
        "semantic_version": "7.2", 
        "versionNumber": 7, 
        "versionMinorNumber": 2, 
        "versionState": "RELEASED"
    }, 
    "dv_link": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/26935&?version=7.2", 
    "ownerInfo": {
        "id": 11, 
        "name": "Gary King Dataverse", 
        "alias": "king", 
        "dv_link": "https://dataverse.harvard.edu/dataverse/king", 
        "dv_link_json": "https://services.dataverse.harvard.edu/miniverse/metrics/v1/dataverses/by-alias/king", 
        "description": null, 
        "dataversetype": "RESEARCHERS"
    }, 
    "metadata_blocks": {
        "citation": {
            "title": "Replication data for: How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It", 
            "dateOfDeposit": "2014-08-04", 
            "distributionDate": "2014", 
            "subject": [
                "Social Sciences"
            ], 
            "distributor": [
                {
                    "distributorName": "Harvard Dataverse", 
                    "distributorLogoURL": "https://dataverse.harvard.edu/resources/images/dataverseproject_logo.jpg"
                }
            ], 
            "publication": [
                {
                    "publicationCitation": "Forthcoming, Political Analysis  <br /><br />  King, Gary, and Margaret Roberts. 2014. How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It. Political Analysis:  <a href=\"http://j.mp/InK5jU\" target=\"_blank\">Link to article</a>"
                }
            ], 
            "dsDescription": [
                {
                    "dsDescriptionValue": "\"Robust standard errors\" are used in a vast array of scholarship to correct standard errors for model misspecification. However, when misspecification is bad enough to make classical and robust standard errors diverge, assuming that it is nevertheless not so bad as to bias everything else requires considerable optimism. And even if the optimism is warranted, settling for a misspecified model, with or without robust standard errors, w\r\nill still bias estimators of all but a few quantities of interest. Even though this message is well known to methodologists, it has failed to reach most applied researchers. The resulting cavernous gap between theory and practice suggests that considerable gains in applied statistics may be possible. We seek to help applied researchers realize these gains via an alternative perspective that offers a productive way to use robust standard errors; a new general and easier-to-use \"generalized information matrix test\" statistic; and practical illustrations via simulations and real examples from published research. Instead of jettisoning this extremely popular tool, as some suggest, we show how robust and classical standard error differences can provide effective clues about model misspecification, likely biases, and a guide to more reliable inferences.\r\n<br /><br /> See also: <a href=\"http://gking.harvard.edu/category/research-interests/methods/unifying-statistical-analysis\" target=\"_blank\">Unifying Statistical Analysis</a>", 
                    "dsDescriptionDate": "2014"
                }
            ], 
            "author": [
                {
                    "authorName": "Roberts, Margaret"
                }, 
                {
                    "authorName": "King, Gary", 
                    "authorAffiliation": "Harvard University"
                }
            ], 
            "datasetContact": [
                {
                    "datasetContactEmail": "king@harvard.edu"
                }
            ]
        }
    }, 
    "files": [
        {
            "id": 2491878, 
            "datasetVersionId": 88878, 
            "name": "bootstrapIM.poisson.R", 
            "description": "", 
            "specs": {
                "storageidentifier": "224179", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 2712, 
                "checksum": {
                    "value": "3e82413d6660f1173f070b4ebe4ec1fc", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491878", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491880, 
            "datasetVersionId": 88878, 
            "name": "bootstrapIM.nb.R", 
            "description": "", 
            "specs": {
                "storageidentifier": "224177", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 3851, 
                "checksum": {
                    "value": "8d5de10d96ffdadafaa39ad00909457e", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491880", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491881, 
            "datasetVersionId": 88878, 
            "name": "Neumayer_ISQ_2003.R", 
            "description": "replicates the Neumayer (2003) example from the paper", 
            "specs": {
                "storageidentifier": "224182", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 8330, 
                "checksum": {
                    "value": "46a79d146977c56da89ad94cbdd4ecd0", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491881", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491882, 
            "datasetVersionId": 88878, 
            "name": "ReadMe.txt", 
            "description": "ReadMe file related to this data", 
            "specs": {
                "storageidentifier": "224188", 
                "contentType": "text/plain; charset=UTF-8", 
                "filesize_bytes": 632, 
                "checksum": {
                    "value": "8446cf916bedcdd873a76b62bcf05b13", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491882", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491883, 
            "datasetVersionId": 88878, 
            "name": "DreherJensen.R", 
            "description": "replicates the Dreher and Jensen (2007) example from the paper", 
            "specs": {
                "storageidentifier": "224181", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 7394, 
                "checksum": {
                    "value": "66c2165ca1a4e639de816f5c0842afe9", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491883", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491884, 
            "datasetVersionId": 88878, 
            "name": "bootstrapIM.normal.R", 
            "description": "", 
            "specs": {
                "storageidentifier": "224178", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 4205, 
                "checksum": {
                    "value": "e02facd1a3ac4c3237b8fbca346a1962", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491884", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491885, 
            "datasetVersionId": 88878, 
            "name": "ButheMilner.R", 
            "description": "replicates the Buthe and Milner (2008) example from the paper.", 
            "specs": {
                "storageidentifier": "224180", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 10841, 
                "checksum": {
                    "value": "dcee264071860da438475f5742c2df2a", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491885", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491887, 
            "datasetVersionId": 88878, 
            "name": "Article for ISQ (aid).tab", 
            "description": "", 
            "specs": {
                "storageidentifier": "224184", 
                "contentType": "text/tab-separated-values", 
                "filesize_bytes": 130826, 
                "checksum": {
                    "value": "0f8534bdf7d31fdacc923e990a55c6b6", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491887", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:24", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491888, 
            "datasetVersionId": 88878, 
            "name": "DreherandJensenJLEreplication.tab", 
            "description": "", 
            "specs": {
                "storageidentifier": "224185", 
                "contentType": "text/tab-separated-values", 
                "filesize_bytes": 321903, 
                "checksum": {
                    "value": "d03a6a8468e1e4a3c0cf3172f08dce12", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491888", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:24", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2493899, 
            "datasetVersionId": 88878, 
            "name": "MilnerButhe2.tab", 
            "description": "Some of the data from the article we replicated by Tim Buthe and Hellen Millner data is proprietary; to obtain it, you must contact them directly.", 
            "specs": {
                "storageidentifier": "226236", 
                "contentType": "text/tab-separated-values", 
                "filesize_bytes": 1267188, 
                "checksum": {
                    "value": "edb06411ed0d6242638163b503919843", 
                    "type": "MD5"
                }
            }, 
            "restricted": true, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2493899", 
            "timestamps": {
                "createdate": "2014-08-05 13:06:13", 
                "publicationdate": "2014-08-04 20:00:00"
            }
        }, 
        {
            "id": 2732877, 
            "datasetVersionId": 88878, 
            "name": "vcovSACCchange.R", 
            "description": null, 
            "specs": {
                "storageidentifier": "151642c1e50-fff554769afe", 
                "contentType": "type/x-r-syntax", 
                "filesize_bytes": 6035, 
                "checksum": {
                    "value": "224a626776b6bf2e955f7cb6ecbefb85", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": "A", 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2732877", 
            "timestamps": {
                "createdate": "2015-12-02 14:29:56", 
                "publicationdate": "2015-12-02 14:31:44"
            }
        }
    ]
}
Older Version
{
    "title": "Replication data for: How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It", 
    "id": 53121, 
    "persistentIdInfo": {
        "protocol": "doi", 
        "authority": "10.7910/DVN", 
        "identifier": "26935", 
        "persistentId": "doi:10.7910/DVN/26935", 
        "persistentUrl": "http://dx.doi.org/10.7910/DVN/26935"
    }, 
    "isPublished": true, 
    "timestamps": {
        "createdate": "2014-08-04 12:44:02", 
        "publicationdate": "2014-08-04 12:52:14"
    }, 
    "semanticVersionInfo": {
        "semantic_version": "5.0", 
        "versionNumber": 5, 
        "versionMinorNumber": 0, 
        "versionState": "RELEASED"
    }, 
    "dv_link": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/26935&?version=5.0", 
    "ownerInfo": {
        "id": 11, 
        "name": "Gary King Dataverse", 
        "alias": "king", 
        "dv_link": "https://dataverse.harvard.edu/dataverse/king", 
        "dv_link_json": "https://services.dataverse.harvard.edu/miniverse/metrics/v1/dataverses/by-alias/king", 
        "description": null, 
        "dataversetype": "RESEARCHERS"
    }, 
    "metadata_blocks": {
        "citation": {
            "datasetContact": [
                {
                    "datasetContactEmail": "king@harvard.edu"
                }
            ], 
            "publication": [
                {
                    "publicationCitation": "King, Gary, and Margaret Roberts. 2014. How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It. Political Analysis:  <a href=\"http://j.mp/InK5jU\" target=\"_blank\">Link to article</a>"
                }
            ], 
            "distributor": [
                {
                    "distributorName": "IQSS Dataverse Network", 
                    "distributorURL": "http://dvn.iq.harvard.edu", 
                    "distributorLogoURL": "http://dvn.iq.harvard.edu/images/iqss-logo-background.png"
                }
            ], 
            "author": [
                {
                    "authorName": "Roberts, Margaret"
                }, 
                {
                    "authorName": "King, Gary", 
                    "authorAffiliation": "Harvard University"
                }
            ], 
            "dsDescription": [
                {
                    "dsDescriptionValue": "\"Robust standard errors\" are used in a vast array of scholarship to correct standard errors for model misspecification. However, when misspecification is bad enough to make classical and robust standard errors diverge, assuming that it is nevertheless not so bad as to bias everything else requires considerable optimism. And even if the optimism is warranted, settling for a misspecified model, with or without robust standard errors, will still bias estimators of all but a few quantities o\nf interest. Even though this message is well known to methodologists, it has failed to reach most applied researchers. The resulting cavernous gap between theory and practice suggests that considerable gains in applied statistics may be possible. We seek to help applied researchers realize these gains via an alternative perspective that offers a productive way to use robust standard errors; a new general and easier-to-use \"generalized information matrix test\" statistic; and practical illustrations via simulations and real examples from published research. Instead of jettisoning this extremely popular tool, as some suggest, we show how robust and classical standard error differences can provide effective clues about model misspecification, likely biases, and a guide to more reliable inferences.\n<br /><br /> See also: <a href=\"http://gking.harvard.edu/category/research-interests/methods/unifying-statistical-analysis\" target=\"_blank\">Unifying Statistical Analysis</a>", 
                    "dsDescriptionDate": "2014"
                }
            ], 
            "title": "Replication data for: How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It", 
            "distributionDate": "2014", 
            "dateOfDeposit": "2014-08-04"
        }
    }, 
    "files": [
        {
            "id": 2491878, 
            "datasetVersionId": 50925, 
            "name": "bootstrapIM.poisson.R", 
            "description": "", 
            "specs": {
                "storageidentifier": "224179", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 2712, 
                "checksum": {
                    "value": "3e82413d6660f1173f070b4ebe4ec1fc", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491878", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491879, 
            "datasetVersionId": 50925, 
            "name": "vcovSACCchange.R", 
            "description": "", 
            "specs": {
                "storageidentifier": "224183", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 5933, 
                "checksum": {
                    "value": "0cf3f65cddf1aa90449f90ad720a8f5a", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491879", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491880, 
            "datasetVersionId": 50925, 
            "name": "bootstrapIM.nb.R", 
            "description": "", 
            "specs": {
                "storageidentifier": "224177", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 3851, 
                "checksum": {
                    "value": "8d5de10d96ffdadafaa39ad00909457e", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491880", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491881, 
            "datasetVersionId": 50925, 
            "name": "Neumayer_ISQ_2003.R", 
            "description": "replicates the Neumayer (2003) example from the paper", 
            "specs": {
                "storageidentifier": "224182", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 8330, 
                "checksum": {
                    "value": "46a79d146977c56da89ad94cbdd4ecd0", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491881", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491882, 
            "datasetVersionId": 50925, 
            "name": "ReadMe.txt", 
            "description": "ReadMe file related to this data", 
            "specs": {
                "storageidentifier": "224188", 
                "contentType": "text/plain; charset=UTF-8", 
                "filesize_bytes": 632, 
                "checksum": {
                    "value": "8446cf916bedcdd873a76b62bcf05b13", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491882", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491883, 
            "datasetVersionId": 50925, 
            "name": "DreherJensen.R", 
            "description": "replicates the Dreher and Jensen (2007) example from the paper", 
            "specs": {
                "storageidentifier": "224181", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 7394, 
                "checksum": {
                    "value": "66c2165ca1a4e639de816f5c0842afe9", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491883", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491884, 
            "datasetVersionId": 50925, 
            "name": "bootstrapIM.normal.R", 
            "description": "", 
            "specs": {
                "storageidentifier": "224178", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 4205, 
                "checksum": {
                    "value": "e02facd1a3ac4c3237b8fbca346a1962", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491884", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491885, 
            "datasetVersionId": 50925, 
            "name": "ButheMilner.R", 
            "description": "replicates the Buthe and Milner (2008) example from the paper.", 
            "specs": {
                "storageidentifier": "224180", 
                "contentType": "text/plain; charset=US-ASCII", 
                "filesize_bytes": 10841, 
                "checksum": {
                    "value": "dcee264071860da438475f5742c2df2a", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491885", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:19", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491887, 
            "datasetVersionId": 50925, 
            "name": "Article for ISQ (aid).tab", 
            "description": "", 
            "specs": {
                "storageidentifier": "224184", 
                "contentType": "text/tab-separated-values", 
                "filesize_bytes": 130826, 
                "checksum": {
                    "value": "0f8534bdf7d31fdacc923e990a55c6b6", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491887", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:24", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2491888, 
            "datasetVersionId": 50925, 
            "name": "DreherandJensenJLEreplication.tab", 
            "description": "", 
            "specs": {
                "storageidentifier": "224185", 
                "contentType": "text/tab-separated-values", 
                "filesize_bytes": 321903, 
                "checksum": {
                    "value": "d03a6a8468e1e4a3c0cf3172f08dce12", 
                    "type": "MD5"
                }
            }, 
            "restricted": false, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2491888", 
            "timestamps": {
                "createdate": "2014-08-04 13:11:24", 
                "publicationdate": "2014-08-03 20:00:00"
            }
        }, 
        {
            "id": 2493899, 
            "datasetVersionId": 50925, 
            "name": "MilnerButhe2.tab", 
            "description": "Some of the data from the article we replicated by Tim Buthe and Hellen Millner data is proprietary; to obtain it, you must contact them directly.", 
            "specs": {
                "storageidentifier": "226236", 
                "contentType": "text/tab-separated-values", 
                "filesize_bytes": 1267188, 
                "checksum": {
                    "value": "edb06411ed0d6242638163b503919843", 
                    "type": "MD5"
                }
            }, 
            "restricted": true, 
            "ingeststatus": null, 
            "file_access_url": "https://dataverse.harvard.edu/api/access/datafile/2493899", 
            "timestamps": {
                "createdate": "2014-08-05 13:06:13", 
                "publicationdate": "2014-08-04 20:00:00"
            }
        }
    ]
}

semanticVersionInfo
semantic_version Modified 5.0 7.2
versionNumber Modified 5 7
versionMinorNumber Modified 0 2

dv_link Modified https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/26935&?version=5.0 https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/26935&?version=7.2

metadata_blocks : citation
subject Added (blank) Social Sciences

metadata_blocks : citation : distributor : Item 1
distributorURL Removed http://dvn.iq.harvard.edu (blank)
distributorName Modified IQSS Dataverse Network Harvard Dataverse
distributorLogoURL Modified http://dvn.iq.harvard.edu/images/iqss-logo-background.png https://dataverse.harvard.edu/resources/images/dataverseproject_logo.jpg

metadata_blocks : citation : publication : Item 1
publicationCitation Modified King, Gary, and Margaret Roberts. 2014. How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It. Political Analysis: Link to article Forthcoming, Political Analysis

King, Gary, and Margaret Roberts. 2014. How Robust Standard Errors Expose Methodological Problems They Do Not Fix, and What to Do About It. Political Analysis: Link to article

metadata_blocks : citation : dsDescription : Item 1
dsDescriptionValue Modified "Robust standard errors" are used in a vast array of scholarship to correct standard errors for model misspecification. However, when misspecification is bad enough to make classical and robust standard errors diverge, assuming that it is nevertheless not so bad as to bias everything else requires considerable optimism. And even if the optimism is warranted, settling for a misspecified model, with or without robust standard errors, will still bias estimators of all but a few quantities o f interest. Even though this message is well known to methodologists, it has failed to reach most applied researchers. The resulting cavernous gap between theory and practice suggests that considerable gains in applied statistics may be possible. We seek to help applied researchers realize these gains via an alternative perspective that offers a productive way to use robust standard errors; a new general and easier-to-use "generalized information matrix test" statistic; and practical illustrations via simulations and real examples from published research. Instead of jettisoning this extremely popular tool, as some suggest, we show how robust and classical standard error differences can provide effective clues about model misspecification, likely biases, and a guide to more reliable inferences.

See also: Unifying Statistical Analysis
"Robust standard errors" are used in a vast array of scholarship to correct standard errors for model misspecification. However, when misspecification is bad enough to make classical and robust standard errors diverge, assuming that it is nevertheless not so bad as to bias everything else requires considerable optimism. And even if the optimism is warranted, settling for a misspecified model, with or without robust standard errors, w ill still bias estimators of all but a few quantities of interest. Even though this message is well known to methodologists, it has failed to reach most applied researchers. The resulting cavernous gap between theory and practice suggests that considerable gains in applied statistics may be possible. We seek to help applied researchers realize these gains via an alternative perspective that offers a productive way to use robust standard errors; a new general and easier-to-use "generalized information matrix test" statistic; and practical illustrations via simulations and real examples from published research. Instead of jettisoning this extremely popular tool, as some suggest, we show how robust and classical standard error differences can provide effective clues about model misspecification, likely biases, and a guide to more reliable inferences.

See also: Unifying Statistical Analysis

Files
File 2732877 Added (blank) id: 2732877
datasetVersionId: 88878
name: vcovSACCchange.R
description: (blank)
specs: storageidentifier: 151642c1e50-fff554769afe
contentType: type/x-r-syntax
filesize_bytes: 6035
checksum: value: 224a626776b6bf2e955f7cb6ecbefb85
type: MD5
restricted: False
ingeststatus: A
file_access_url: https://dataverse.harvard.edu/api/access/datafile/2732877
timestamps: createdate: 2015-12-02 14:29:56
publicationdate: 2015-12-02 14:31:44
File 2491879 Removed id: 2491879
datasetVersionId: 50925
name: vcovSACCchange.R
description:
specs: storageidentifier: 224183
contentType: text/plain; charset=US-ASCII
filesize_bytes: 5933
checksum: value: 0cf3f65cddf1aa90449f90ad720a8f5a
type: MD5
restricted: False
ingeststatus: (blank)
file_access_url: https://dataverse.harvard.edu/api/access/datafile/2491879
timestamps: createdate: 2014-08-04 13:11:19
publicationdate: 2014-08-03 20:00:00
(blank)