Common use of Choosing PRPL Records for Your Research Question Clause in Contracts

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 or HC-039. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 2000? Select the Panel 5 round 3 and Panel 4 round 5 records with PRIVCAT>0 and MSUPINS ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 2000? Select the Panel 5 round 3 and Panel 4 round 5 records with DNTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 1 and Panel 4 round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s record. Merge each record onto PUF HC-022 and use weights, strata, and PSUs to create nationally representative estimates.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-159 or HC-039HC-165. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three four research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002014? Select the Panel 5 round 19 Round 3 and Panel 4 round 18 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-165 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002014? Select the Panel 5 round 19 Round 3 and Panel 4 round 18 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,51, 4, 5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-165 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 19 Round 1 and Panel 4 round 18 Round 3 records with PRIVCAT in (2, 3, 6, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-159 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2014, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 19 Round 1 or Panel 18 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2014 JOBS file, PUF HC-166, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 18 may have reported retiring from a job in 2013, so, from the 2013 JOBS file, PUF HC-158, select the records with PANEL=18 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-165 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-143 or HC-039HC-149. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three five research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002012? Select the Panel 5 round 17 Round 3 and Panel 4 round 16 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-149 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002012? Select the Panel 5 round 17 Round 3 and Panel 4 round 16 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,51, 4, 5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-149 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 17 Round 1 and Panel 4 round 16 Round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-143 and use weights, strata, and PSUs to create nationally representative estimates. How many people were in families that gave the highest rating for at least one of their private health plans? Select the Panel 17 Round 2 and Panel 16 Round 4 records with SATELIG=1 and RATEPLAN=10. Select one record for each DUPERSID. Merge each record onto PUF HC-149 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2012, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 17 Round 1 or Panel 16 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2012 JOBS file, PUF HC-150, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 16 may have reported retiring from a job in 2011, so, from the 2011 JOBS file, PUF HC-142, select the records with PANEL=16 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-149 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-117 or HC-039HC-123. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three five research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002009? Select the Panel 5 round 14 Round 3 and Panel 4 round 13 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-123 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002009? Select the Panel 5 round 14 Round 3 and Panel 4 round 13 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-123 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 14 Round 1 and Panel 4 round 13 Round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-117 and use weights, strata, and PSUs to create nationally representative estimates. How many people were in families that gave the highest rating for at least one of their private health plans? Select the Panel 14 Round 2 and Panel 13 Round 4 records with SATELIG=1 and RATEPLAN=10. Select one record for each DUPERSID. Merge each record onto PUF HC-123 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2009, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 14 Round 1 or Panel 13 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2009 JOBS file, PUF HC-124, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 13 may have reported retiring from a job in 2008, so, from the 2008 JOBS file, PUF HC-116, select the records with PANEL=13 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIREJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-123 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-151 or HC-039HC-157. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three four research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002013? Select the Panel 5 round 18 Round 3 and Panel 4 round 17 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-157 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002013? Select the Panel 5 round 18 Round 3 and Panel 4 round 17 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,51, 4, 5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-157 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 18 Round 1 and Panel 4 round 17 Round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-151 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2013, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 18 Round 1 or Panel 17 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2013 JOBS file, PUF HC-158, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 17 may have reported retiring from a job in 2012, so, from the 2012 JOBS file, PUF HC-150, select the records with PANEL=17 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-157 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-084 or HC-039HC-090. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three five research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002005? Select the Panel 5 round 10 Round 3 and Panel 4 round 9 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-090 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002005? Select the Panel 5 round 10 Round 3 and Panel 4 round 9 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-090 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 10 Round 1 and Panel 4 round 9 Round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-084 and use weights, strata, and PSUs to create nationally representative estimates. How many people were in families that gave the highest rating for at least one of their private health plans? Select the Panel 10 Round 2 and Panel 9 Round 4 records with SATELIG=1 and RATEPLAN=10. Select one record for each DUPERSID. Merge each record onto PUF HC-090 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2005, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 10 Round 1 or Panel 9 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2005 JOBS file, PUF HC-091, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 9 may have reported retiring from a job in 2004, so, from the 2004 JOBS file, PUF HC-083, select the records with PANEL=9 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIREJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-090 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-053 or HC-039HC-062. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002002? Select the Panel 5 7 round 3 and Panel 4 6 round 5 records with PRIVCAT>0 and MSUPINS ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-062 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002002? Select the Panel 5 7 round 3 and Panel 4 6 round 5 records with DNTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-062 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 7 round 1 and Panel 4 6 round 3 records with PRIVCAT XXXXXXX in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s record. Merge each record onto PUF HC-022 HC-053 and use weights, strata, and PSUs to create nationally representative estimates.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-013 or HC-039HC-038. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three four research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20001999? Select the Panel 5 4 round 3 and Panel 4 3 round 5 records with PRIVCAT>0 and MSUPINS ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-038 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20001999? Select the Panel 5 4 round 3 and Panel 4 3 round 5 records with DNTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-038 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 4 round 1 and Panel 4 round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s record. Merge each record onto PUF HC-022 HC-013 and use weights, strata, and PSUs to create nationally representative estimates. How many people who had private health insurance for hospitals/physicians in January 1998 also had health insurance in December 1999 from the same source? From the 1999 Person Round Plan file (HC-047), select all the panel 3 round 1 records with PRIVCAT>0 and MSUPINS ≠ 1 and STATUS1=1. Select the DUPERSID and EPRSIDX variables. Then select all the round 5 records with STATUS24=1, and select the EPRSIDX and DUPERSID variables. A successful match between round 1 and round 5 records is coverage from the same source in both times. A person might, however, have coverage from more than one source in both January and December, so to avoid double counting persons, eliminate any duplicate DUPERSIDs before merging the round 1 and round 5 records. Merge the matches onto PUF HC-035 and use longitudinal weights, strata, and PSUs to create nationally representative estimates.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-125 or HC-039HC-132. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three five research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002010? Select the Panel 5 round 15 Round 3 and Panel 4 round 14 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-132 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002010? Select the Panel 5 round 15 Round 3 and Panel 4 round 14 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-132 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 15 Round 1 and Panel 4 round 14 Round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-125 and use weights, strata, and PSUs to create nationally representative estimates. How many people were in families that gave the highest rating for at least one of their private health plans? Select the Panel 15 Round 2 and Panel 14 Round 4 records with SATELIG=1 and RATEPLAN=10. Select one record for each DUPERSID. Merge each record onto PUF HC-132 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2010, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 15 Round 1 or Panel 14 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2010 JOBS file, PUF HC-133, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 14 may have reported retiring from a job in 2009, so, from the 2009 JOBS file, PUF HC-124, select the records with PANEL=14 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-132 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-167 or HC-039HC-174. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three four research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002015? Select the Panel 5 round 20 Round 3 and Panel 4 round 19 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-174 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002015? Select the Panel 5 round 20 Round 3 and Panel 4 round 19 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,51, 4, 5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-174 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 20 Round 1 and Panel 4 round 19 Round 3 records with PRIVCAT in (2, 3, 6, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-167 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2015, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 20 Round 1 or Panel 19 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2015 JOBS file, PUF HC-176, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 19 may have reported retiring from a job in 2014, so, from the 2014 JOBS file, PUF HC-166, select the records with PANEL=19 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-174 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-064 or HC-039HC-073. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three five research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002003? Select the Panel 5 8 round 3 and Panel 4 7 round 5 records with PRIVCAT>0 and MSUPINS ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-073 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002003? Select the Panel 5 8 round 3 and Panel 4 7 round 5 records with DNTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-073 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 8 round 1 and Panel 4 7 round 3 records with PRIVCAT XXXXXXX in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s 's record. Merge each record onto PUF HC-022 HC-064 and use weights, strata, and PSUs to create nationally representative estimates. How many people were in families that gave the highest rating for at least one of their private health plans? Select the Panel 8 round 2 and Panel 7 round 4 records with SATELIG=1 and RATEPLAN=10. Select one record for each DUPERSID. Merge each record onto PUF HC-073 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2003, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 8 round 1 or Panel 7 round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2003 JOBS file, PUF HC-074, select the records with for jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 6 may have reported retiring from a job in 2001, so, from the 2002 JOBS file, PUF HC-063, select the records with PANEL=6 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIREJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-073 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-101 or HC-039HC-107. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three five research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002007? Select the Panel 5 round 12 Round 3 and Panel 4 round 11 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-107 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002007? Select the Panel 5 round 12 Round 3 and Panel 4 round 11 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-107 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 12 Round 1 and Panel 4 round 11 Round 3 records with PRIVCAT in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-101 and use weights, strata, and PSUs to create nationally representative estimates. How many people were in families that gave the highest rating for at least one of their private health plans? Select the Panel 12 Round 2 and Panel 11 Round 4 records with SATELIG=1 and RATEPLAN=10. Select one record for each DUPERSID. Merge each record onto PUF HC-107 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2007, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 12 Round 1 or Panel 11 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2007 JOBS file, PUF HC-108, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 11 may have reported retiring from a job in 2006, so, from the 2006 JOBS file, PUF HC-100, select the records with PANEL=11 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIREJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-107 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-177 or HC-039HC-184. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three four research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002016? Select the Panel 5 round 21 Round 3 and Panel 4 round 20 Round 5 records with PRIVCAT>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-184 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002016? Select the Panel 5 round 21 Round 3 and Panel 4 round 20 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT in (1,4,51, 4, 5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-184 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 21 Round 1 and Panel 4 round 20 Round 3 records with PRIVCAT in (2, 3, 6, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-177 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2016, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 21 Round 1 or Panel 20 Round 3) and PHOLDER=1 and PRIVCAT=1 and EVALCOVR=1]. From the 2016 JOBS file, PUF HC-185, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Persons in Panel 20 may have reported retiring from a job in 2015, so, from the 2015 JOBS file, PUF HC-176, select the records with PANEL=20 and (SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIRJOB=1 or YLEFT=2 or YNOBUSN=2 or WHY_LEFT=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-184 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. ‌ In order to produce estimates from the data in this file, researchers must use the person (person-level or family) -level weights released in either of two previously released PUFs, HC-022 HC-205 or HC-039HC-212. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three four research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002019? Select the Panel 5 round 24 Round 3 and Panel 4 round 23 Round 5 records with PRIVCAT>0 PrivateCat>0 and MSUPINS MSUPINSX ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-212 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002019? Select the Panel 5 round 24 Round 3 and Panel 4 round 23 Round 5 records with DNTLINS=1 DENTLINS=1 and PRIVCAT PrivateCat in (1,4,51, 4) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-212 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 round 24 Round 1 and Panel 4 round 23 Round 3 records with PRIVCAT PrivateCat in (2, 3, 995, 6) and MSUPINSX ne 1 and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC-205 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview of 2019, how many people had insurance from jobs from which they retired? Select the PRPL records for policyholders of employment-related insurance at the time of the first interview [(Panel 24 Round 1 or Panel 23 Round 3) and PHOLDER=1 and PrivateCat =1 and EVALCOVR=1]. From the 2019 JOBS file, PUF HC-211, using JOBSIDX, select the records with jobs from which the person retired (SUBTYPE=6 or RETIRJOB=1 or YNOBUSN_M18=2 or WHY_LEFT_M18=3). Persons in Panel 23 may have reported retiring from a job in 2018, so, from the 2018 JOBS file, PUF HC-203, using JOBSIDX select the records with PANEL=23 and (SUBTYPE=6 or RETIRJOB=1 or YNOBUSN=2 or WHY_LEFT=3). Combine the records from the two JOBS files, keeping only one record per JOBSIDX. Using the JOBSIDX, merge the selected JOBS records onto the selected PRPL records. Select the PRPL records with SUBTYPE=6 or RETIRJOB=1 or YNOBUSN_M18=2 or WHY_LEFT_M18=3 or EMPLSTAT=2. Select one record for each DUPERSID. Merge each record onto PUF HC-212 and use weights, strata, and PSUs to create nationally representative estimates of the number of people with one of these PRPL records.

Choosing PRPL Records for Your Research Question. In order to produce estimates from the data in this file, researchers must use the person (or family) level weights released in either of two previously released PUFs, HC-022 HC-034 or HC-039HC-055. Researchers must consult the documentation for these PUFs for guidance on creating nationally representative estimates for different time periods. Note that if there are multiple records per person (DUPERSID) when you merge on weights, you will double count some people, and your estimates will not be nationally representative. There are two solutions: select only one record per person, or aggregate information across PRPL records. How you develop your analytical file depends on your research question. The PRPL file is designed to help answer a wide variety of research questions. AHRQ cannot anticipate all these questions, so this section provides examples of how to use the PRPL file for three research questions. How many people were covered by two or more private hospital/physician insurance plans at the end of 20002001? Select the Panel 5 6 round 3 and Panel 4 5 round 5 records with PRIVCAT>0 and MSUPINS ne 1 and EVALCOVR=1. Count the number of records for each person (DUPERSID). Create one person- person-level record for each DUPERSID that has the number of plans (PRPL records). Merge the count variable onto PUF HC-039 HC-055 and use weights, strata, and PSUs to create nationally representative estimates. How many people reported private dental coverage from an employer at the end of 20002001? Select the Panel 5 6 round 3 and Panel 4 5 round 5 records with DNTLINS=1 and PRIVCAT in (1,4,5) and EVALCOVR=1. Among these records, select one record for each person (DUPERSID). Merge each record onto PUF HC-039 HC-055 and use weights, strata, and PSUs to create nationally representative estimates. At the time of the first interview, how many private insurance policies for hospital/physician were not employment-based? Select the Panel 5 6 round 1 and Panel 4 5 round 3 records with PRIVCAT XXXXXXX in (2, 3, 99) and EVALCOVR=1. Select one record for each policyholder-establishment pair (EPRSIDX). To have a positive weight for the final count, we recommend choosing the covered person record of the policyholder (PHOLDER=1), unless the policyholder is deceased (DECPHLDR=1), in which case then the researcher should choose a different covered person=s ’s record. Merge each record onto PUF HC-022 HC- 034 and use weights, strata, and PSUs to create nationally representative estimates.