Announcement

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

  • Help Needed: Simple Linear Regression and Significance Questions

    Hi Everyone!

    I'm relatively new to STATA and statistics, so I thought I would come to this forum for some help.

    I'm working with data that relates different factors to the returns on schooling. Some of these factors include race, number of siblings, IQ, birthdate, location, years of work experience, years of schooling, etc.

    I need to plot the total returns to experience and calculate when it is maximized.
    My data set has two variables for experience:
    experience and experiencesquared

    Should I use one variable instead of the other? How would I go about plotting the total returns and calculating when it is maximized?

    PS. If anyone knows how to conduct a test of the overall significance of a regression, please let me know!

    Thank you for your help and please let me know if there is any other information I should provide

  • #2
    Welcome to the Stata Forum / Statalist.

    I gather by ‘regression’ you meant - regress - command (please read the FAQ and find extra explanation).

    Under - regress - you have the ‘omnibus’ F test as well as the R-square, to name a few items to look for.

    With regard to plotting results, I strongly recommend to take a look at - margins - and - marginsplot- commands.

    To end, depending on the distribution pattern of returns, you may consider using a generalized linear model. For example, if returns are highly skewed, a gamma family should stay on the radar.

    Hopefully that helps.
    Best regards,

    Marcos

    Comment


    • #3
      Sarah:
      if your question related to a graded assigment, please see https://www.statalist.org/forums/help#adviceextras #4.
      That said, a regression requires one predictor only (which is not your case); in all likelihood, you are planning to run a multiple regression.
      Being obscure the regressand (is return to school a binary variable, ie no=0 and yes=1?),I find difficult to advise any further in this respect.
      As far as the squared term is concerned, see -fvvarlist- notation instead of plugging in the right-hand side of your regerssion equation two variables (the first linear, the second squared) for the same predictor (-experience-). Please note that -fvvarlist- notation only allows you to exploit in full the wonderful capabilities of -margins- and -marginsplot- that Marcos wisely recommended for your analysis.
      Kind regards,
      Carlo
      (StataNow 18.5)

      Comment

      Working...
      X