Individuals surveyed in only one outgoing rotation group?

I am interested in analyzing union membership for people who transition between employment and nonemployment in the CPS ORG.

I have noticed that there appear to be many people who appear in month 4 but not month 8 or month 8 but not month 4 of the survey.

I did a quick count of how many times each cpsidp appears in months 4 or 8. The data I am using are from January 2010 to December 2019, but this could be easily looked at in a smaller sample.

keep if inlist(mish,4,8) & eligorg == 1
duplicates tag cpsidp, generate(dups)

  count_obs |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |    618,893       39.75       39.75
          2 |    937,906       60.25      100.00
------------+-----------------------------------
      Total |  1,556,799      100.00

From this, it looks like about 40 percent of the sample is in month 4 or month 8 (or neither) but not both. I know that people who left the survey in 2010 or people who first entered the survey in 2019 will not be included twice.

When I look at the date of these individuals only surveyed once, I see the following:

       date |      Freq.     Percent        Cum.
------------+-----------------------------------
     2010m1 |      7,899        1.58        1.58
     2010m2 |      8,225        1.64        3.22
     2010m3 |      8,018        1.60        4.82
     2010m4 |      8,486        1.69        6.51
     2010m5 |      7,996        1.60        8.10
     2010m6 |      8,450        1.69        9.79
     2010m7 |      8,221        1.64       11.43
     2010m8 |      8,246        1.64       13.07
     2010m9 |      8,272        1.65       14.72
    2010m10 |      8,337        1.66       16.39
    2010m11 |      7,984        1.59       17.98
    2010m12 |      8,301        1.66       19.64
     2011m1 |      2,829        0.56       20.20
     2011m2 |      2,862        0.57       20.77
     2011m3 |      2,875        0.57       21.34
     2011m4 |      2,868        0.57       21.92
     2011m5 |      2,798        0.56       22.47
     2011m6 |      2,884        0.58       23.05
     2011m7 |      2,752        0.55       23.60
     2011m8 |      2,830        0.56       24.16
     2011m9 |      2,780        0.55       24.72
    2011m10 |      2,968        0.59       25.31
    2011m11 |      2,927        0.58       25.89
    2011m12 |      2,963        0.59       26.48
     2012m1 |      2,778        0.55       27.04
     2012m2 |      2,844        0.57       27.61
     2012m3 |      2,888        0.58       28.18
     2012m4 |      2,895        0.58       28.76
     2012m5 |      2,776        0.55       29.31
     2012m6 |      2,850        0.57       29.88
     2012m7 |      2,971        0.59       30.47
     2012m8 |      2,875        0.57       31.05
     2012m9 |      2,977        0.59       31.64
    2012m10 |      2,951        0.59       32.23
    2012m11 |      3,003        0.60       32.83
    2012m12 |      3,026        0.60       33.43
     2013m1 |      2,868        0.57       34.01
     2013m2 |      2,961        0.59       34.60
     2013m3 |      2,892        0.58       35.17
     2013m4 |      2,970        0.59       35.77
     2013m5 |      2,940        0.59       36.35
     2013m6 |      2,884        0.58       36.93
     2013m7 |      3,004        0.60       37.53
     2013m8 |      2,961        0.59       38.12
     2013m9 |      2,876        0.57       38.69
    2013m10 |      3,043        0.61       39.30
    2013m11 |      3,739        0.75       40.04
    2013m12 |      3,651        0.73       40.77
     2014m1 |      3,426        0.68       41.46
     2014m2 |      3,762        0.75       42.21
     2014m3 |      3,932        0.78       42.99
     2014m4 |      3,762        0.75       43.74
     2014m5 |      3,606        0.72       44.46
     2014m6 |      3,790        0.76       45.22
     2014m7 |      4,000        0.80       46.01
     2014m8 |      3,942        0.79       46.80
     2014m9 |      3,783        0.75       47.55
    2014m10 |      4,059        0.81       48.36
    2014m11 |      4,179        0.83       49.20
    2014m12 |      4,215        0.84       50.04
     2015m1 |      4,095        0.82       50.86
     2015m2 |      4,069        0.81       51.67
     2015m3 |      3,975        0.79       52.46
     2015m4 |      4,201        0.84       53.30
     2015m5 |      4,061        0.81       54.11
     2015m6 |      4,119        0.82       54.93
     2015m7 |      3,792        0.76       55.69
     2015m8 |      3,834        0.76       56.45
     2015m9 |      3,580        0.71       57.17
    2015m10 |      3,950        0.79       57.95
    2015m11 |      3,071        0.61       58.57
    2015m12 |      3,182        0.63       59.20
     2016m1 |      3,284        0.66       59.86
     2016m2 |      3,218        0.64       60.50
     2016m3 |      3,146        0.63       61.12
     2016m4 |      3,242        0.65       61.77
     2016m5 |      3,285        0.66       62.43
     2016m6 |      3,335        0.67       63.09
     2016m7 |      3,382        0.67       63.77
     2016m8 |      3,461        0.69       64.46
     2016m9 |      3,404        0.68       65.14
    2016m10 |      3,288        0.66       65.79
    2016m11 |      3,388        0.68       66.47
    2016m12 |      3,401        0.68       67.15
     2017m1 |      3,223        0.64       67.79
     2017m2 |      3,251        0.65       68.44
     2017m3 |      3,418        0.68       69.12
     2017m4 |      3,132        0.62       69.74
     2017m5 |      3,152        0.63       70.37
     2017m6 |      3,075        0.61       70.99
     2017m7 |      3,261        0.65       71.64
     2017m8 |      2,940        0.59       72.22
     2017m9 |      3,206        0.64       72.86
    2017m10 |      3,070        0.61       73.48
    2017m11 |      3,260        0.65       74.13
    2017m12 |      3,123        0.62       74.75
     2018m1 |      3,163        0.63       75.38
     2018m2 |      3,090        0.62       76.00
     2018m3 |      3,181        0.63       76.63
     2018m4 |      3,113        0.62       77.25
     2018m5 |      3,177        0.63       77.89
     2018m6 |      3,178        0.63       78.52
     2018m7 |      3,204        0.64       79.16
     2018m8 |      3,018        0.60       79.76
     2018m9 |      3,195        0.64       80.40
    2018m10 |      3,051        0.61       81.01
    2018m11 |      3,165        0.63       81.64
    2018m12 |      3,092        0.62       82.25
     2019m1 |      7,634        1.52       83.78
     2019m2 |      7,336        1.46       85.24
     2019m3 |      7,537        1.50       86.74
     2019m4 |      7,438        1.48       88.23
     2019m5 |      7,386        1.47       89.70
     2019m6 |      7,541        1.50       91.21
     2019m7 |      7,165        1.43       92.63
     2019m8 |      7,311        1.46       94.09
     2019m9 |      7,348        1.47       95.56
    2019m10 |      7,419        1.48       97.04
    2019m11 |      7,469        1.49       98.53
    2019m12 |      7,378        1.47      100.00
------------+-----------------------------------
      Total |    501,313      100.00

There appear to be a few thousand people each month in years 2011-2018 who were surveyed only in month 4 or 8 of their rotation group but not both.

Am I doing something wrong with the data, or are these individuals surveyed in only one of their outgoing rotation groups?

@Duncan_Hobbs

Thanks for asking this question - I encountered the same issue so I am interested in the answer as well.

(My explanation was to attribute it to attrition but this does not square well with being in month 8 only.)

The trend you are seeing is consistent with what is expected for CPS basic monthly data. This is because of non-response, mortality, migration (households drop out and are replaced by the households that move into their housing unit), and recording errors, which lead to non-trivial amounts of single-response data. More information can be found on the description tab for CPSIDP.