Announcement

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

  • Survival Time for cumulative employment spells

    I am trying to create a variable for "Years of Employment Experience" for use in a panel data set.
    My data are individuals observed of a period of 18 years, 1991 to 2009.

    I have data of when each employment spell started (year), and when it ended (year) for each individual (identified by an ID number variable) as well as the year the individual was interviewed (in the main data set, which is separate).

    I have managed to add up total employment spells across the individual's lifetime so far, but need to transfer this to the panel data.

    I am try to create a ‘cumulative experience’ variable so that the spells of employment are only summed up to the year of interview.

    Can this be done using survival time commands, or if not, what would be the best way to proceed?

    Any help would be much appreciated.

    Many thanks!

    Ella



    Last edited by Ella Ki; 23 Mar 2017, 21:13.

  • #2
    The survival time commands are for doing survival analysis. You're still at the point of setting up your data.

    This is not a difficult problem, but your description of your data is not detailed enough to enable anyone to provide a specific solution. Rather than trying to describe it better in words, which will be quite lengthy and probably hard to follow, you should show an example of your data.

    In order to make the example you show usable to somebody who wants to help you (because that person will probably want to test his/her solution code on your data before posting it), you need to use the -dataex- command. Run -ssc install dataex- to install the command and then run -help dataex- and read the simple instructions for using it. Create a compact example of your data, probably based on one or two representative individuals and including all of the relevant variables, and use -dataex- to post it here on the Forum. You will probably get a helpful response in a short time if you do that.

    Comment


    • #3
      Originally posted by Clyde Schechter View Post
      The survival time commands are for doing survival analysis. You're still at the point of setting up your data.

      This is not a difficult problem, but your description of your data is not detailed enough to enable anyone to provide a specific solution. Rather than trying to describe it better in words, which will be quite lengthy and probably hard to follow, you should show an example of your data.

      In order to make the example you show usable to somebody who wants to help you (because that person will probably want to test his/her solution code on your data before posting it), you need to use the -dataex- command. Run -ssc install dataex- to install the command and then run -help dataex- and read the simple instructions for using it. Create a compact example of your data, probably based on one or two representative individuals and including all of the relevant variables, and use -dataex- to post it here on the Forum. You will probably get a helpful response in a short time if you do that.
      Thanks a lot for the advice - I am new to the Forum.

      My main data is from the British Household Panel Survey and extra employment spells data was created by a researcher from the BHPS.

      Since I don't own the data, would it be alright for me to do that?

      Thanks again!

      Comment


      • #4
        Since I don't own the data, would it be alright for me to do that?
        I don't know. You would have to consult the documentation and instructions that came with your survey. If you aren't allowed to disseminate the actual data, what really matters for your question is the data layout and the variable names. So you could just take your data set, and keep the first few dozen observations. Then keep only the ID numbers and the survey years and the start and end dates. If the ID numbers themselves might be used to identify real people (e.g. they are social security numbers or something like that), just change them to consecutive numbers 1, 2, 3, 4, ... etc (of course always using the same new 1, 2, 3, 4 number for the same original ID number). Then post just the (changed) IDs and the dates. The other variables that are in the surveys are probably not relevant to answering your specific question.

        Comment

        Working...
        X