Announcement

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

  • Generalized ordered logit model and clarification about thresholds

    Hi,

    I found a paper that estimates a generalized ordered logit model. The dependent variable is wealth quintiles (1,2,3,4,5) and it controls for other variables as well (e.g income quintiles, age etc).

    My confusion concerns the way the authors have named wealth quintiles. They state wealth quintiles from 1st to 4th but shouldn't it be named wealth thresholds?

    I have also attached a picture of the relevant table.

    Thanks
    Ilias
    Attached Files
    Last edited by Ilias Geo; 11 Jan 2019, 05:54.

  • #2
    Think about quartiles. There are four segments of data that they divide the data into, but there are only three of them (25th, 50th and 75th percentiles).

    So a quantile is not a segment of data, but rather the threshold as you call it, and there is one fewer quantile than the number of segments of data. (Clearest in median.)

    Would this help?

    Comment


    • #3
      Yes, thank you

      Comment

      Working...
      X