Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • How to produce Odds ratio stata output to latex

    Dear Users,


    I needed to produce stata out put to latex. The out put is the odds ratio of logit estimation.
    I am using the following command :
    outreg2 using "file name", keep(var) stats(coef se) eform bdec(3) sdec(3) nocons label

    The command produced the odds ratio and the standard errors associated with the odds ratio. So, i have two questions.

    1. Should i report these standard errors that are associated with the odds ratio or should i report the standard errors from the original logit estimation?
    2. I know the odds ratio are e the power of the coefficients from the logit estimation, how are the standard errors associated with the odds ratio computed?


    Best,
    Last edited by Habtamu Beshir; 05 Sep 2018, 00:42.

  • #2
    1) If you want to report the standard errors then I would report the standard errors for the odds ratios. This is an approximation using the delta method, which assumes that the sampling distribution of the odds ratio can be approximated using a normal distribution. This will often work fine, but not always. Safer is to report the z-statistic for the raw coefficient. That is the test statistic I would use for testing whether the odds ratio equals 1 (or, equivalently, the log odds ratio equals 0). See: https://www.stata.com/support/faqs/s...cs/delta-rule/

    2) As mentioned above, the delta method is used to compute the standard errors. If the raw coefficient (log odds ratio) is called \(b\), its standard error \(se_b\) , and the standard error of the odds ratio \(se_{or}\), then \(se_{or}=\exp(b)se_b\)
    ---------------------------------
    Maarten L. Buis
    University of Konstanz
    Department of history and sociology
    box 40
    78457 Konstanz
    Germany
    http://www.maartenbuis.nl
    ---------------------------------

    Comment


    • #3
      Thanks so much, Maarten.
      Very helpful.

      Best,

      Comment

      Working...
      X