{
  "abstract": "This special feature contains some personal suggestions for improving quantitative analyses. I focus on articles that use regressions to determine the connections among variables, a common way for evaluating the effects of public policy. Examples come from the firearms literature. It seems to me that, once they have obtained some data, too many researchers almost immediately start running regressions—before they fully understand the dataset or think deeply about the questions they are trying to answer. I provide six suggestions for researchers: Try to: (1) determine and report a causal theory, including the chain-of-causation; (2) investigate the accuracy of the data; (3) explore the data; (4) disaggregate where possible; (5) determine if the results are plausible and (6) be transparent about the methods and results—let the reader into your ‘statistical kitchen.’ The evidence about the effectiveness of Child Access Prevention Laws illustrates some of these issues.",
  "authors": [
    {
      "affiliations": [
        "Harvard TH Chan School of Public Health Harvard Injury Control Research Center, Boston, Massachusetts, USA"
      ],
      "name": "David Hemenway"
    }
  ],
  "title": "Six suggestions for improving quantitative evaluations",
  "uid": "de0e9f11-dc5f-51ac-869c-f30d0194ac6e"
}