Understanding Society User Support

Hi,

As far as I understand, the employ variable simply flags if someone is employed or not. And I think the nunmpsp_dv variable is the number of unemployment spells since the last interview. The description for nunmpsp_dc (https://www.understandingsociety.ac.uk/documentation/mainstage/dataset-documentation/variable/nunmpsp_dv) states that "In this case, when either last year's employment status or current year's employment status was unequal to unemployed, it is assumed that a spell in "unemployment' was completed."

I have ran some code which identifies if someone's 'employ' status changed from one wave to the next, and then cross-tabulated this against the nunmpsp_dv variable. The two variables don't seem to match up at all. There are many observations where someone was unemployed at either the last wave or the current wave according to 'employ', but who have 0 unemployment spells since last wave according to nunmpsp_dv.

The cross tab is below for waves g - h of the usoc. The columns are derived from employ, and the rows are nunmpsp_dv.

move into work move out of work remain employed remain unemployed
  -7              0              143               0              1161
  -8            350                6              23               287
  -9              2                2              12                 5
  0            1127             1174           17307             11741
  1             354              282             123              1195
  2               9                1               1                31
  3               2                0               2                 1

Please could you explain why this might be the case? Let me know if you need me to share any code.

Many thanks,
Mike