U.S. flag

An official website of the United States government

Skip Header


2014 Panel Wave 1: Recode allocation flags

The 2008 SIPP panel offered widely used allocation flags for the earnings from each job or business:  APMSUM1/2 and ABMSUM1/2.  Given that we constructed the earnings “recode” variables TPMSUM1/2 and TBMSUM1/2 by combining responses to multiple instrument fields, it is not clear how one should construct allocation flag values for these variables.  In constructing APMSUM1/2 and ABMSUM1/2, we made various decisions which do not necessarily align with decisions that a particular data user would have made.  Moreover, the process by which we constructed TPMSUM1/2 and TBMSUM1/2 was rather complicated, yielding uncertainty in the informational content of APMSUM1/2 and ABMSUM1/2.

Recognizing these shortcomings of APMSUM1/2 and ABMSUM1/2 in the 2008 panel, we have endeavored to be more transparent in constructing allocation flags for earnings recodes in the 2014 panel.  Although flags are present on the 2014 SIPP panel data file for our earnings recode variables, these allocation flags indicate only that data users should construct their own allocation flags for the earnings recode variables using the allocation flags for all of the variables that serve as inputs to these recode variables.  Because we recognize that users will want access to more meaningful allocation flags for earnings recode variables, we have written sample code to construct allocation flags for TJB(n)_MSUM (for n=1,…,7) and TPEARN.  Interested parties can find this code at the website of the Census Bureau’s Labor Force Statistics Branch.

This sample code will clarify to users which decisions must be made in order to construct AJB(n)_MSUM and APEARN.  The code is written to be used without modification by the most conservative data user possible.  In brief, AJB(n)_MSUM indicates imputation if any variable that determines TJB(n)_MSUM is imputed.  More specifically:

  • If EJB(n)_SCRNR was imputed to a value of 2, indicating the sample member did not hold job (n), then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed.  If not,
    • If EJB(n)_STARTWK or EJB(n)_ENDWK is imputed or if EJB(n)_STARTWK and EJB(n)_ENDWK is out of universe but EJB(n)_BMONTH or EJB(n)_EMONTH is imputed in some other month of the year, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed.  If not,
    • If EJB(n)_AWOP(i), EJB(n)_AWOPSW(i), or EJB(n)_AWOPEW(i) are imputed for any away without pay spell for i=1,…,3, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed.  If not,
    • If EJB(n)_TYPPAY1, EJB(n)_INCPB, or EJB(n)_BSLRYB are imputed, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed.  If not,
    • If EJB(n)_CHEARN1, EJB(n)_CHEARN2, EJB(n)_CHERWK1, or EJB(n)_CHERWK2 are imputed, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed.  If not,
    • If week XWKB(j) falls during pay spell i and if EJB(n)_ANNSAL(i), EJB(n)_HOURLY(i), EJB(n)_WKLY(i), EJB(n)_BWKLY(i), EJB(n)_MTHLY(i), EJB(n)_SMTHLY(i), EJB(n)_OTHER(i), or EJB(n)_GAMT1 are imputed, then wage and salary earnings is assumed to be imputed for week XWKB(j).  If EJB(n)_PAYHR(i)=1, then we also refer to EJB(n)_AWOPCH(k) for k=1,…,3.  If any of these variables are imputed, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed for week XWKB(j).  If EJB(n)_PAYHR(i)=2, then we also refer to EJB(n)_CHHOUR1/2, EJB(n)_CHHOWK1/2, and EJB(n)_JOBHRS1/2/3.  If EJB(n)_CHHOUR1, EJB(n)_CHHOUR2, EJB(n)_CHHOWK1, or EJB(n)_CHHOWK2 are imputed, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed for week XWKB(j).  If not, and if week XWK(j) falls during hours of work spell m and if EJB(n)_JOBHRS(m) is imputed, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed for week XWKB(j).  Finally, if wage and salary earnings during any XWKB(j) in a particular month are imputed, then the contribution of wage and salary earnings to TJB(n)_MSUM is assumed to be imputed.
  • If EJB(n)_BMONTH or EJB(n)_EMONTH is imputed, or if EJB(n)_BMONTH and EJB(n)_BMONTH are out of universe but EJB(n)_BMONTH or EJB(n)_EMONTH is imputed in some other month of the year, then the contribution of extra earnings (including tips, commission, overtime, and bonus payments, and other income from self-employed jobs) to TJB(n)_MSUM is assumed to be imputed.  If not,
    • If EJB(n)_TYPPAY2, EJB(n)_TYPPAY3, EJB(n)_TYPPAY4, EJB(n)_TYPPAY5, EJB(n)_OINCB, EJB(n)_INCPB, or EJB(n)_BSLRYB are imputed, then the contribution of extra earnings to TJB(n)_MSUM is assumed to be imputed.  If not,
    • If EJB(n)_BXAMT, EJB(n)_CXAMT, EJB(n)_OXAMT, EJB(n)_TXAMT, or EJB(n)_OINCAMT are imputed, then the contribution of extra earnings to TJB(n)_MSUM is assumed to be imputed.
  • If either the contribution of wage and salary earnings to TJB(n)_MSUM is imputed or the contribution of extra earnings to TJB(n)_MSUM is imputed, then TJB(n)_MSUM is assumed to be imputed.
  • If EJB(n)_JBORSE was imputed to a value other than 2, then the contribution of profits to TPEARN is assumed to be imputed.  If not,
    • If EJB(n)_STARTWK or EJB(n)_ENDWK is imputed or if EJB(n)_STARTWK and EJB(n)_ENDWK is out of universe but EJB(n)_BMONTH or EJB(n)_EMONTH is imputed in some other month of the year, then the contribution of profits to TPEARN is assumed to be imputed.  If not,
    • If EJB(n)_AWOP(i), EJB(n)_AWOPSW(i), or EJB(n)_AWOPEW(i) are imputed for any away without pay spell for i=1,…,3, then the contribution of profits to TPEARN is assumed to be imputed.  If not,
    • If EJB(n)_PFTLOSS is imputed, then the contribution of profits to TPEARN is assumed to be imputed.  If not,
    • If TJB(n)_PRFTB is imputed, then the contribution of profits to TPEARN is assumed to be imputed.
  • If TJB(n)_MSaUM is imputed for any job n=1,…,8, or if the contribution of profits to TPEARN is imputed for any job n=1,…,7, then TPEARN is assumed to be imputed.

Data users who prefer to make less conservative decisions about how to construct earnings recode allocation flags can comment out lines of code or add new code as needed.

Page Last Revised - October 8, 2021
Is this page helpful?
Thumbs Up Image Yes Thumbs Down Image No
NO THANKS
255 characters maximum 255 characters maximum reached
Thank you for your feedback.
Comments or suggestions?

Top

Back to Header