Announcement

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

  • Propensity score and matching for specific year in balanced panel data set

    Dear Statalisters,

    As part of my thesis I am conducting a DiD analysis on a balanced data set with US families from 2007 to 2021. In that matter, I want to match the treatment and control group based on income, marital status, and age (covariates). However, after searching the internet I am not sure how to do this with, as the treatment is defined on behalf of their wealth in 2017 before a reform, but I also have a dummy indicating 1 for all years for each family if they are treated. But I am afraid that Stata counts 1 for all years and creates he propensity score for each year. What I want is for each family to have only one propensity score from 2017. I want to match the data only on behalf of their treatment status in 2017. I have made a dummy indicating if they are treated in 2017 or not. My data is in long format.
    So, the DiD is made from 2007 to 2021, with treatment status defined in 2017 just before the reform.

    The outcome variable is whether or not they own a house (ownrate) and I tend to use nearest neighbour matching.


    Code:
    YEAR    HOUSEHOLDID    MARITAL    treatment    treatmentall    ownrate    avage    loginc
    2017    1066    Married    0    0    0    44.5    11.46163
    2017    1019    Never married    0    0    1    37    11.27643
    2015    3225    Divorced, annulled    0    0    0    67    9.174506
    2017    1122    Divorced, annulled    0    0    0    51    11.46528
    2007    2486    Never married    0    0    0    34    9.563529
    2011    669    Divorced, annulled    0    0    0    54    11.27973
    2005    3175    Married    0    0    0    35.5    11.24114
    2011    2551    Married    0    0    1    54.5    10.18998
    2007    876    Married    0    0    1    53.5    11.75194
    2013    1707    Never married    0    0    0    34    9.746834
    2007    109    Divorced, annulled    0    0    0    35    9.973806
    2019    3094    Married    0    0    0    72.5    10.45057
    2011    1194    Never married    0    0    0    27    9.753827
    2011    2834    Married    0    0    0    52    10.60287
    2011    3094    Married    0    0    1    64.5    11.05232
    2019    3209    Never married    0    0    0    57    8.864323
    2013    1095    Married    0    0    1    49    10.87805
    2011    2308    Married    0    0    1    55.5    10.63164
    2009    2639    Separated    0    0    1    60    9.10498
    2005    1681    Married    0    0    1    32    11.04541
    2007    2326    Separated    0    0    1    48    8.919454
    2017    1815    Never married    0    0    0    38    12.76719
    2009    3238    Never married    0    0    0    29    8.517193
    2013    2771    Never married    0    0    1    66    10.2485
    2015    2827    Divorced, annulled    0    0    0    35    10.22818
    2013    3132    Never married    0    0    0    51    9.021598
    2017    2923    Married    0    0    1    61    11.33675
    2009    3034    Widowed    0    0    0    70    8.881836
    2017    2771    Never married    0    0    1    70    10.75332
    2011    7    Married    0    0    0    33    10.23996
    2021    1245    Never married    0    0    1    59    
    2011    3132    Never married    0    0    0    49    8.998137
    2005    2830    Never married    0    0    0    47    
    2007    1109    Widowed    0    0    1    63    10.10643
    2015    13    Never married    0    0    0    49    8.34284
    2011    1128    Divorced, annulled    0    0    0    40    9.814875
    2013    935    Married    0    0    0    59    10.98597
    2017    921    Widowed    0    0    1    69    11.32662
    2009    806    Married    0    0    0    35    11.28978
    2005    3275    Never married    0    0    0    23    10.00785
    2017    2182    Widowed    0    0    0    93    9.103645
    2007    3047    Separated    0    0    1    30    10.30535
    2019    3226    Never married    0    0    0    46    9.73696
    2015    1765    Divorced, annulled    0    0    0    49    11.00227
    2019    2920    Married    0    0    0    43.5    11.28643
    2021    1508    Married    0    0    1    47.5    
    2013    206    Married    0    0    1    60    10.93521
    2015    839    Never married    0    0    0    29    9.082052
    2015    2453    Never married    0    0    0    41    8.268732
    2021    3068    Widowed    0    0    1    69    
    2011    410    Married    0    0    1    51.5    11.759
    2013    165    Married    0    0    0    48    10.68281
    2005    1088    Married    0    0    1    32    10.34174
    2015    1448    Married    0    0    0    70.5    10.96992
    2017    3043    Divorced, annulled    0    0    0    61    10.63219
    2007    886    Never married    0    0    1    27    11.25156
    2021    2109    Married    0    0    1    39    
    2021    2028    Never married    0    0    0    42    
    2015    3235    Married    0    0    0    41    10.59663
    2013    2284    Married    0    0    1    49.5    9.908176
    2007    2701    Married    0    0    0    50    10.60162
    2019    2876    Divorced, annulled    0    0    0    61    10.55059
    2021    474    Married    0    0    1    45.5    
    2021    1128    Divorced, annulled    0    0    0    50    
    2007    3130    Divorced, annulled    0    0    0    50    11.0021
    2011    3206    Married    0    0    0    45.5    11.48839
    2011    699    Married    0    0    1    68    11.13411
    2015    3122    Divorced, annulled    0    0    1    48    9.789535
    2021    489    Separated    0    0    1    47    
    2009    1805    Divorced, annulled    0    0    0    47    8.738735
    2011    2550    Never married    0    0    0    72    9.763765
    2015    2849    Divorced, annulled    0    0    0    44    11.351
    2019    372    Married    0    0    1    63.5    10.93756
    2017    165    Married    0    0    0    52.5    11.28978
    2021    1045    Never married    0    0    0    40    
    2009    3364    Never married    0    0    0    58    11.21167
    2005    691    Married    0    0    0    27    10.94553
    2011    39    Married    0    0    1    54    10.07036
    2007    1367    Never married    0    0    0    27    10.81978
    2021    1640    Married    0    0    1    51    
    2015    561    Never married    0    0    0    73    10.4445
    2017    1411    Married    0    0    1    32.5    11.54217
    2009    3229    Married    0    0    1    47    11.55215
    2005    123    Married    0    0    1    49.5    10.59663
    2013    2684    Never married    0    0    0    48    
    2019    472    Married    0    0    1    59    10.63826
    2015    2098    Never married    0    0    0    50    9.472705
    2013    2828    Never married    0    0    0    32    7.824046
    2021    3349    Never married    0    0    0    47    
    2013    2049    Divorced, annulled    0    0    0    64    10.08643
    2009    2958    Widowed    0    0    0    27    8.131531
    2021    1100    Married    0    0    1    69    
    2015    3243    Divorced, annulled    0    0    1    68    10.05818
    2005    2098    Never married    0    0    0    40    9.903487
    2013    362    Never married    0    0    0    47    10.71914
    2009    2609    Divorced, annulled    0    0    0    39    10.91696
    2013    3328    Widowed    0    0    0    85    9.421249
    2013    191    Married    0    0    1    37    11.05645
    2017    2486    Never married    0    0    0    45    10.70759
    2015    1154    Never married    0    0    0    46    11.00225
    2009    2302    Married    0    0    0    45.5    10.37349
    2005    2681    Married    0    0    0    39.5    11.2051
    2005    1550    Divorced, annulled    0    0    0    50    10.54534
    2013    2138    Married    0    0    0    37.5    10.78808
    2021    3109    Widowed    0    0    1    80    
    2021    2618    Divorced, annulled    0    0    0    59    
    2013    3395    Never married    0    0    1    54    9.560716
    2017    3065    Divorced, annulled    0    0    0    46    10.75594
    2021    1246    Married    0    0    0    50.5    
    2015    2546    Never married    0    0    0    34    
    2009    2588    Never married    0    0    0    54    10.40426
    I really hope someone can help.

  • #2
    Per default, matching is not a panel method. A simple solution is as follows: match in the year of interest (cross sectionally), generate weights and use these weights in your subsequent analyses (perhaps even a panel model). Test with caution.
    Best wishes

    (Stata 16.1 MP)

    Comment


    • #3
      Thank you. So you suggest, that at match on a sub sample from 2017, then find the families with a match, and continue with the those in the original data set?

      Comment


      • #4
        Yes if the treatment was applied in 2017 this is the relevant data for the matching application.
        Best wishes

        (Stata 16.1 MP)

        Comment


        • #5
          Thank you very much. This was super helpful!

          Comment


          • #6
            If you go along with matching be sure to use kmatch by Ben Jann.
            Best wishes

            (Stata 16.1 MP)

            Comment

            Working...
            X