Announcement

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

  • Svyset for pooled cross-sectional data set

    Hello all,

    I have a pooled data set of three waves of the Afrobarometer surveys that I'm struggling to declare as survey data using the svyset command. The three survey ways have already been appended and according to the survey manuals, the PSU/EA is the urban/rural placement while stratification is usually provinces divided into urban and rural areas. I have the respondent's id variable, country variable, the region variable, the urbrur variable, and two weight variables from the data sets.

    The weight variables account for the following:
    withinwt: within country weighting factor
    Combinwt: multi-country weighting factor.

    My question is: how do I correctly subset such data? I will be doing cross-sectional analysis where individuals are nested in countries. Hopefully I'll get help regarding this.

    My current code for this is:

    pweight: Combinwt
    VCE: linearized
    Single unit: missing
    Strata 1: urbrur
    SU 1: regions
    FPC 1: <zero>

    I also tried a bivariate regression with the following code:

    urvey: Linear regression

    Number of strata = 2 Number of obs = 68,031
    Number of PSUs = 1,032 Population size = 53,114.663
    Design df = 1,030
    F( 1, 1030) = 299.29
    Prob > F = 0.0000
    R-squared = 0.0251

    ------------------------------------------------------------------------------
    | Linearized
    swd | Coef. Std. Err. t P>|t| [95% Conf. Interval]
    -------------+----------------------------------------------------------------
    ownecon | .1326844 .0076697 17.30 0.000 .1176345 .1477343
    _cons | 2.124502 .0300417 70.72 0.000 2.065552 2.183452
    ------------------------------------------------------------------------------


    Based on the examples above, have I done the svyset correctly?

    Thank you in advance for your help.

    Regards,
    Rita


  • #2
    Survey: Linear regression

    Number of strata = 2 Number of obs = 68,031
    Number of PSUs = 1,032 Population size = 53,114.663
    Design df = 1,030
    F( 1, 1030) = 299.29
    Prob > F = 0.0000
    R-squared = 0.0251


    Linearized
    swd Coef. Std. Err. t P>t [95% Conf. Interval]

    ownecon .1326844 .0076697 17.30 0.000 .1176345 .1477343
    _cons 2.124502 .0300417 70.72 0.000 2.065552 2.183452

    Comment


    • #3
      Rita, Could you tell me if you were able to verify the svy weights for pooled Afrobarometer surveys? I am trying to figure out how do the the same.
      Best wishes, Mark

      Comment

      Working...
      X