Announcement

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

  • Interpretation of coefficients in LPM

    Dear all,

    Perhaps a very basic question, but all the online resources I have looked at so far give me a different answer to the following question.

    In a linear probability model, how should the coefficient on a dummy independent variable be interpreted? For instance, say we have the model

    Yi=a+b1Male+b2 Xi+ui

    where Y is 1 if the individual participates in the labour market, 0 if not and Male is a dummy that's 1 if the individual is male, 0 if female. Let's say the coefficient b1 is 0.10. Does this mean that being a man increases the probability to be in the labour force by 10 percent or by 10 percentage points (compared to being a woman)?

    I think it should be 10 percentage points, but in many lectures, they speak of 10% .

    Kind regards!

  • #2
    Percent and percentage points are the same thing. Basis points are 1/100 of a percent, which is what you may be thinking of. In any case, both interpretations are correct, b1=0.1 would be a 10% increase in the probability of being a labor force participant if you are a male.

    Comment


    • #3
      No, percent and percentage point are not the same. I have been called on this many times! For example, see

      http://thewritingbusiness.com/the-di...entage-points/
      -------------------------------------------
      Richard Williams, Notre Dame Dept of Sociology
      StataNow Version: 18.5 MP (2 processor)

      EMAIL: [email protected]
      WWW: https://www3.nd.edu/~rwilliam

      Comment


      • #4
        It is generally a mistake to refer to probabilities as if they were percentages, for exactly the reasons demonstrated here. I would instead say that a coefficient of 0.10 means that being a man increases the probability of being in the labor force by 0.10 compared to being a woman.

        Comment


        • #5
          Richard Williams Thank you, this was exactly the difference I had in mind! So am I correct in thinking that the coefficient should be interpreted as the percentage points change in probability of participation if the dummy equals 1?

          Comment


          • #6
            This handout explains how to interpret the lpm and why the lpm is problematic. https://www3.nd.edu/~rwilliam/xsoc73994/Logit01.pdf
            -------------------------------------------
            Richard Williams, Notre Dame Dept of Sociology
            StataNow Version: 18.5 MP (2 processor)

            EMAIL: [email protected]
            WWW: https://www3.nd.edu/~rwilliam

            Comment


            • #7
              Richard Williams William Lisowski Thank you both!

              Comment


              • #8
                Percentage points is correct. An example I give in my book is when people in the state of Michigan were asked to vote on increasing the sales tax rate from 4 percent to 6 percent. Those in favor referred to it as a two percentage point increase in the tax rate. Those opposed said it was a 50 percent increase in the tax rate. Both were correct.

                Comment


                • #9
                  https://www.explainxkcd.com/wiki/index.php/985:_Percentage_Points

                  Click image for larger version

Name:	percentage_points.png
Views:	2
Size:	32.4 KB
ID:	1512869
                  David Radwin
                  Senior Researcher, California Competes
                  californiacompetes.org
                  Pronouns: He/Him

                  Comment

                  Working...
                  X