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

  • interpret margin effects after logit regressions

    I want to interpret the economic significance of the margin effects. I have the following commands, I want to examine the association between satisfaction (satisfy - a continuous variable) and probability of leaving the firm (leave).

    logit leave satisfy other_controls

    margins, dydx (satisfy) atmeans post

    | Delta-method
    | dy/dx Std. Err. z P>|z| [95% Conf. Interval]
    satisfy | .0137787 .0050354 2.74 0.006 .0039095 .0236479

    Variable | Obs Mean Std. Dev. Min Max
    | 15,593 -.1700978 .7704771 -2.17395 2.455053

    Variable | Obs Mean Std. Dev. Min Max
    leave| 15,593 .2015648 .4011817 0 1

    In terms of the economic significance, did I interpret it correctly?
    Holding other variables at means,one unit decrease in satisfaction score, increases employees' likelihood to leave the firm by
    1.37 percentage points.
    Holding other variables at means, a one standard deviation decrease in the rating of satisfaction will increase employees' propensity of departure by about 5.2% of the mean(0.0137*0.77)/0.202

    Thank you very much in advance.


  • #2
    Hi Ava. Your post would be easier to read if you used code tags. See pt 12 of the Statalist FAQ on asking Qs effectively.

    No, the interpretation is not quite right. How close an approximation it is depends on the scaling of satisfy. See

    I know economists seem to love marginal effects of continuous variables, but I don't. I prefer plotting adjusted predictions at different values of X. The mcp command from SSC is very good for this. See
    Richard Williams, Notre Dame Dept of Sociology
    Stata Version: 16.1MP (2 processor)

    EMAIL: rwilliam@ND.Edu


    • #3
      Thanks very much, Richard.