| 2024 | 2023 | 2022 | 2021 | 2020 | 2019 | 2018 |
October 31, 2024
- IPUMS MEPS now includes data from the 2022 sample; over 1,100 variables from the Full Year Consolidated, Condition, Event, Prescribed Medicine, Appendix to the Event, and Pooled Variance files are now available. New variables include information about COVID and Long COVID (CVDEVER, CVDLONG, CVDLONGNOW, CVDLONGACT) and INTERVTYPRD, which indicates the method by which AHRQ administered the questionnaire in each round. Users should be aware that the original 2022 conditions data unexpectely included 7,655 conditions with no corresponding medical event that took place during the 2022 calendar year. Since 2018, the MEPS conditions data have included only treated conditions, or those conditions associated with a medical event during the calendar year. We are monitoring the AHRQ website for additional documentation or guidance about these non-adherent cases and will share guidance when we learn more.
- The round-level variable CVDVAX has been re-named CVDVAXRD to align with the naming conventions of other Round variables.
- INCWELFR is now available for sample year 2021. It was inadvertently excluded from the 2021 release.
- An error in the variable USCOWNGRP incorrectly reported responses of "Unknown - don't know" as "Unknown - not ascertained." This has been corrected.
- In this release, we also corrected a long-standing error in the variable INSCOP, "Inscope status for the portion of Round 5/3 that covers the current year." We had previously offered the original MEPS variable INSCOPE ("Whether person was ever inscope during year") as INSCOP, which is incorrect. The IPUMS variable INSCOP now correctly reflects the original MEPS variable INSCOP (inscope status for the portion of round 5/3 that covers the current year) and the variable description has been updated with accurate information.
New data.
Changed data.
Corrected data.
July 19, 2024
- The code for NIU cases in MULTC1, MULTC1S1 and MULTC1S1S1 have been changed from 8 to 998.
- The User Note on Analysis and Variance Estimation has been updated to include guidance on sampling weight and survey design specification for analysis of longitudinal data.
- The variable code page now displays code frequencies for the following variables MULTC1, MULTC1S1, MULTC1S1S1, MULTC1S1S2, MULTC1S2, MULTC1S3, MULTC1S3S1, MULTC1S2S1, CCCODE, CCRS1, CCRS2, CCRS3, ICD9CODE, and ICD10CODE.
- Updates have been made to the documentation on the following variables: RXDRGNAME, MULTC1, MULTC1S1 and MULTC1S1S1 to alert users to AHRQ corrections.
Modified data.
Modified documentation.
October 4, 2023
- IPUMS MEPS now includes 2021 data; over 900 variables (both annual person-level variables and round-level variables) from the 2021 file are now available through the extract system. Variables new to this year include COVID variables (CVDBOOSTER and CVDVAX) and variables from the Social Determinants of Health (SDOH) supplement (see description of SATISFIED for full list of variables).
- The width of HHAGCHTOT has increased from 6 to 7 columns wide to accomodate larger values.
- Version D071.V2.2 of IPUMS MEPS was skipped due to production delays related to changes made to accommodate the 2020 extended MEPS panel. See the IPUMS Health Surveys DOI project page for a complete listing of IPUMS MEPS DOI versions.
New Data.
Changed data.
DOI Note.
July 27, 2023
- The width of STRATANN has increased from 4 to 8 columns wide to accommodate the addition of YEAR as a prefix. Efforts to reproduce MEPS published estimates on the AHRQ data tools website indicated that we were able to perfectly reproduce point estimates but only able to imperfectly reproduce standard errors when using the value of strata directly from the annual AHRQ files in data sets combining two or more years. By adding YEAR as a prefix to the value of strata from the annual files, we are able to perfectly reproduce published point estimates and standard errors from the AHRQ data tools website. For more information, see the IPUMS MEPS note on Analysis and Variance Estimation with MEPS.
- IPUMS MEPS now includes records derived from the Prescribed Medicines file. These include over 70 new variables across three new record types comprised of Medication-Round Fill (F), Medication-Round (M), and Event-Medication-Round link (K) records collected on the MEPS since 1996. F records appear and can be added to extracts under the RX Medicines menu. M record variables are automatically added when an F record variable is selected and K record variables are automatically added when an Event (E) record variable is selected.
Changed data.
New Data.
April 21, 2023
- IPUMS MEPS now includes 2020 FYC data; over 900 variables (both annual person-level variables and round-level variables) from the 2020 file are now available through the extract system. Variables new to this year include COVID variables (CVDLAYCA, CVDLAYDN, CVDLAYPM) and a wage outlier flag OUTFLAG. The MEPS interview structure and sample design were adjusted to accommodate nonresponse bias due to the COVID-19 pandemic. Face-to-face interviews were replaced with telephone interviews and Panel 23 was extended to nine rounds (rather than the standard five), which led to changes in how weights were developed for FY20. For more information, see the MEPS Full Year 2020 user note.
New Data.
Modified Data.
June 9, 2022
- IPUMS MEPS has integrated the pooled variance variables (PSUPLD and STRATAPLD) for 2016-2019, and added integrated versions of the round-level variables PREGNTRD, MOMPIDRD, and POPPIDRD and a handful of event-level variables: LABTEST, SONOGRAM, XRAYS, MAMMOGRAM, EKG, DRSPLTY, MEDPTYPE, GENDENT, EXAMINE, NUMNIGHX, EMERROOM, RSNINHOS, DLVRTYPE, EPIDURAL, HHTYPE, CNA, COMPANN, HHAIDE, HOSPICE, HHDAYSMO, and SELFAGEN.
New Data.
May 2, 2022
- IPUMS MEPS now includes records derived from the Medical Condition and Event files. These records include over 100 new variables collected on the MEPS since 1996. For more information, please see our overview of the event and medical conditions variables, including how those variables are offered through IPUMS MEPS, and general considerations to keep in mind when conducting analyses using these variables.
New Data.
Corrected data.
January 24, 2022
- IPUMS MEPS staff reviewed the set of variables from the Access to Care section of MEPS, finalized the variable documentation accordingly, and introduced the new variables YRSINUSC, USCOWNGRP, DELAYDTCST, and DELAYRXCST. The Access to Care section gathers information on usual source of care for all family members, characteristics of usual source of care health providers, and barriers family members have faced in obtaining needed health care. Please see ACCESSELIG for an overview of the available variables.
- Previously, for the 1996 sample, USHOSYNONE was incorrectly based on the original MEPS variable that NOUSLYNONE is based on. Additionally, DIFCARFMYEQP, DIFCARFMYWRK, and DIFCARFMYWHR were incorrectly based on the original MEPS variable that DIFCARFMYIND is based on. The 1996 sample is now using the correct original MEPS variables.
- The codes for the "yes" and "no" values for the 1996 sample of USHOSYDK were accidentally flipped. The error has been corrected.
- Previously, the codes for the "Unknown-not ascertained" and "Unknown-don't know" values for the 1996 sample of USCCHNG were flipped. The error has been corrected.
- The variable YRSINUS was renamed to YRSINUSG. Additionally, the labels for the "unknown" categories have been changed. They were previously "7 - Refused," "8 - Don't know", and "9 - Not ascertained." They are now "7 - Unknown-refused," "8 - Unknown-not ascertained," and "9 - Unknown-don't know."
- The label for the value of 7 in WHYNOUSLPL was previously "No place available in area" and has been updated to "Usual source of care provider not available in area." Additionally, the code for "Other reason" was changed from 30 to 40. Previously, for 1996-2001, the codes for "Other insurance-related reason" was incorrectly mapped to "No health insurance." The error has been corrected. Finally, the code 12 was changed from "No health insurance" to "Usual source of care doctor is somewhere else." "No health insurance" now takes the code 32.
- To accommodate greater detail in later samples, the codes for TYPUSCTRANS have been expanded from a width of 1 to 2. Previously, the codes were "0 - NIU," "1 - Drive/Is driven," "2 - Taxi, bus, train, other public transport," "3 - Walks," "4 - Other way," "7 - Unknown-refused," "8 - Unknown-not ascertained," and "9 - Unknown-don't know." Now, the codes are "0 - NIU," "10 - Drives or is driven," "11 - Drives," "12 - Is driven," "20 - Taxi, bus, train, other public transport," "30 - Walks," "40 - Other way," "97 - Unknown-refused," "98 - Unknown-not ascertained," and "99 - Unknown-don't know."
- The values for USUALPLFMY were adjusted to accommodate the addition of a new code. The codes 1-7 and 97-99 remain unchanged. Previously, the codes were "8 - Seldom sick/don't need usual source," "9 - Other insurance-related reason," and "10-Other." Now, the codes are "8 - Cost-related reason," "9 - Seldom sick/don't need usual source," "10 - Other insurance-related reason," and "11 - Other."
- Previously, the 2013 sample for USCPRLANG only included responses from individuals in Panel 18. It now includes the repsonses of those in both Panel 17 and Panel 18.
Modified documentation.
Corrected data.
Modified data.
December 9, 2021
- IPUMS MEPS now includes 75 new variables from the Child Preventive Health Supplement. For more information see LESSHLTHY.
- Previously the following Self-Administered Questionnaire (SAQ) variables were only available for adults ages 18 and older: ADRTCR, ADRTWW, ADILCR, ADILWW, VISITYRNO, ADNDCR, ADEGMC, ADLIST, ADEXPL, ADRESP, ADPRTM, ADHECR, ADSPEC, ADESSP. With the addition of the Child Preventive Health Supplement, these variables are now available for adults ages 18 and older and children ages 0 to 17.
- Previously there was a mistake in 2012 for ADHECR (Rating of all health care received: last 12 months). A rating of 1 was coded as 10, a rating of 2 was coded as 1, a rating of 3 was coded as 2, a rating of 4 was coded as 3, a rating of 5 was coded as 4, a rating of 6 was coded as 5, a rating of 7 was coded as 6, a rating of 8 was coded as 7, a rating of 9 was coded as 8, and a rating of 10 was coded as 9. This has been corrected so that a rating of 1 is coded as 1, a rating of 2 is coded as 2, etc.
New Data.
Modified Data.
Corrected Data.
October 26, 2021
- Previously, in hierarchical extracts, the values for the record types "person" and "round" appeared as "2" and "3," respectively. Now, the record types appear as "P" and "R."
Modified data.
September 14, 2021
- IPUMS MEPS now includes 2019 FYC data; over 900 variables (both annual person-level variables and round-level variables) are now available through the extract system.
- The variable CPI2009 previously reported the incorrect CPI conversion factor for 2016-2018. Additionally, IPUMS NHIS staff updated the values of CPI2009 to use the most recently published statistics by the Federal Reserve Board. These changes are minimal and should have no substantive effect on analaysis. The errors have been corrected and modifications have been made both in the data and the corresponding user note.
- Previously, for the 2018 MEPS sample DIAEYEXCALYR erroneously referred to the MEPS variable DSEY1753, which has a reference period of the previous calendar year. DIAEYEXCALYR now correctly refers to the original MEPS variable for having a dilated eye exam during the current calendar year.
- Previously, for the 2018 MEPS sample DIAFTCKCALYR erroneously referred to the MEPS variable DSFT1753, which has a reference period of the previous calendar year. DIAFTCKCALYR now correctly refers to the original MEPS variable for having feet checked during the current calendar year.
- Previously, for the 2018 MEPS sample DIACHLCALYR erroneously referred to the MEPS variable DSCH1753, which has a reference period of the previous calendar year. DIACHLCALYR now correctly refers to the original MEPS variable for having blood cholesterol checked during the current calendar year.
- Previously, for the 2018 MEPS sample DIAFLCALYR erroneously referred to the MEPS variable DSFL1753, which has a reference period of the previous calendar year. DIAFLCALYR now correctly refers to the original MEPS variable for receiving a flu vaccination during the current calendar year.
- DEPRESS2WK has been renamed to PHQDEP to match the IPUMS NHIS names of a substantively similar variable that was introduced as part of the 2019 NHIS sample.
- The codes for TYPPLSICK were changed and expanded to be three digits wide. The categories of TYPPLSICK were previously: 00 = "NIU"; 10 = "Clinic or health center"; 11 = "Health center"; 12 = "Public health clinic"; 13 = "Clinic, unspecified type"; 14 = "Community/immigrant/rural clinic"; 15 = "Company or school clinic/center"; 16 = "Urgent care/walk-in clinic"; 17 = "Migrant clinic"; 18 = "Rural health center"; 19 = "Community or family health center"; 20 = "Doctor's office or HMO"; 21 = "Doctor's office"; 22 = "HMO"; 23 = "Doctor's clinic"; 24 = "Private doctor's office in hospital"; 30 = "Hospital emergency room"; 40 = "Hospital outpatient department"; 41 = "County/city/public/county hospital outpatient"; 42 = "Private/other hospital outpatient clinic"; 50 = "Some other place"; 51 = "Psychiatric hospital/clinic"; 52 = "VA hospital/clinic"; 53 = "Military health care facility"; 54 = "Military or VA health care facility"; 55 = "At home"; 56 = "Non-hospital place"; 57 = "Other places (1985: Includes known HMOs)"; 60 = "Doesn't go to one place most often"; 97 = "Unknown-refused"; 98 = "Unknown-not ascertained"; and 99 = "Unknown-don't know."
- Codes for EDUC have been updated to accommodate a change in the coding of the underlying education data introduced in the 2019 NHIS data, which collapsed grades 1-11 into a single category. The coding structure for MEPS is the same as previous years, but the changes to the NHIS coding structure required significant revising of existing categories. Previously, the categories of EDUC were: 100 = "Grade 8 or less"; 101 = Never attended/kindergarten only; 102 = Grade 1; 103 = Grade 2; 104 = Grade 3; 105 = Grade 4; 106 = Grade 5; 107 = Grade 6; 108 = Grade 7; 109 = Grade 8; 200 = Grade 9-12, no diploma; 201 = Grade 9; 202 = Grade 10; 203 = Grade 11; 204 = 12th grade, no diploma; 300 = High school diploma or GED; 301 = High school graduate; 302 = GED or equivalent; 400 = Some college, no 4 yr degree; 401 = Some college, no degree; 402 = AA degree: technical/vocational/occupational; 403 = AA degree: academic program; 500 = Bachelor's degree (BA,AB,BS,BBA); 600 = Master's, Professional, or Doctoral Degree; 601 = Master's degree (MA,MS,Med,MBA); 602 = Professional degree (MD,DDS,DVM,JD); 603 = Doctoral degree (PhD,EdD); 604 = Other degree; 996 = No degree, years of education unknown; 997 = Unknown--refused; 998 = Unknown -- not ascertained; 999 = Unknown -- don't know.
New data.
Corrected data.
Modified data.
October 8, 2020
- IPUMS MEPS now includes 2018 FYC data; over 900 variables (both annual person-level variables and round-level variables) are now available through the extract system.
- ADAPPT, ADDPRS, and ADEZUN have been renamed to VISITYRNO, DEPRESS2WK, and EZINFOFREQ, respectively, to match the IPUMS NHIS names of substantively similar variables.
- MEPSID and MEPSIDRD have taken on a new meaning. Beginning in 2018, AHRQ began offering a unique record identifier on the original MEPS files as a combination of PANEL + DUPERSID. IPUMS MEPS staff created this variable for the 1996-2017 samples and now offer this identifier as MEPSID and MEPSIDRD. MEPSID2 and MEPSID2RD now take on the old meaning of the IPUMS-created MEPSID as a combination of DUID + PID +PANEL, but these variables will not be offered from 2018 forward.
- To accommodate the introduction of a new code in the original data, the code for "Other" was changed from 10 to 11 for WHYNOWK2RD.
- On extracts containing round data structured as wide instead of long, the negative values of FTOTINCCPS, FTOTINCMEPS, INCTOT, FTOTVAL, INCBUS, INCPROP, INCTRST, POVLEV, and MCS were incorrectly written out as missing values on all data formats except CSV. The issue has been corrected and negative values are now being correctly handled.
New data.
Modified data.
Corrected data.
June 16, 2020
- IPUMS MEPS now includes the full set of cancer diagnosis, age of diagnosis, and remission status variables. For more information, see CNBLAD. The variables AGELAST, indicating a respondent's age the last time they were eligible for MEPS, and HINOTCOV, an indicator for respondents with no health insurance, are also now available.
- CPI2009 was updated to reflect the 2017 CPI U adjustment factor for the 2017 sample. Previously, all values were zero. Additionally, we corrected an error in the codes for the variable USCPRFAC. "Person" was erroneously coded as "facility" and "facility" coded as "person." This error has been corrected.
- The coding for the 1999 sample of ACCESSELIG was updated to increase comparability across years. The cases that were previously coded to "Not in universe" are now coded to "No."
- The universe for USCPRFAC was incorrect. It previously indicated that persons without a usual source of care provider (USUALPL) were asked this question. The universe was updated to properly reflect that all persons who do have a usual source of care (USUALPL) were asked this question.
- The universe for USCPRTYP has been updated to state that persons with a usual source of care provider (USUALPL) who is either a person, a person in a facility, or was not ascertained are in universe for USCPRTYP. The universe previously did not include persons whose usual source of care provider is a person in a facility.
- Similarly, the universes for USCPRASN, USCPRBLK, USCPRHISP, USCPRNATAM, USCPROPI, USCPROTHRAC, USCPRWHT, and USCPRSEX have been updated to state that all persons with a usual source of care provider who is a person or a person in a facility are in universe. Previously, only persons with a usual source of care provider who is a person were listed as in universe.
New data.
Corrected data.
Modified data.
Corrected documentation.
February 19, 2020
- IPUMS MEPS now includes data from the Diabetes Care Survey (DCS) from 2000 to the present. For a full list of the variables included in the supplement, please see DIACONF. COVERTYPE and COVTYPEAGE, which measure health insurance coverage type and health insurance coverage type by age, respectively, have also been added to IPUMS MEPS.
- The following "master" variables previously indicated in error that their respective associated expenditure and monthly health insurance variables were available for the 2017 MEPS sample: AMPAEXSRC, AMCHIEXSRC, AMOPTEXSRC, AMNPEXSRC, AMTHEXSRC, DVGEXSRC, DVOEXSRC, OBPAEXSRC, OBCHIEXSRC, OBOPTEXSRC, OBNPEXSRC, OBNPHEXSRC, OBTHEXSRC, OBNPHEXSRC, OPNPHPREXSRC, STAMM, ZNPREXSRC, ZNFEXSRC. The associated variables were added to the data cart when users selected the master variable, but these variables would have had no data for the 2017 sample. This error has been fixed and the availability information listed in each variable's webpage is now correct.
- The codes for the variable RULETRRD in 2016 were updated. Although RULETRRD still uniquely identified a Reporting Unit within a Dwelling Unit in 2016, we updated values of "X" to "B" for greater consistency with other years of data.
New data.
Corrected data.
Modified data.
December 3, 2019
Modified data.
October 3, 2019
- We corrected an error in the values of PSUANN in 2002. Previously, PSUANN contained the values of STRATANN by mistake.
Corrected data.
September 25, 2019
- IPUMS MEPS now includes 2017 FYC data; over 900 variables (both annual person-level variables and round-level variables) are now available through the extract system.
- The codes for the variable ADPRXY were updated. The new codes accommodate the more general coding structure used by AHRQ in 2017. Please see the codes page for more information.
- The width for the variable HIEU was changed from 12 to 7 wide after the elimination of right-side Bs in 1999 and 2000 only. HIEU previously did and continues to function as a unique identifier.
- We fixed an error that incorrectly collapsed family identifiers in the 1996 and 2012 data for the variable FAMIDRD; family identifier values of "C" and "D" were combined into a single category ("C") and, in 2012 only, values of "E" in the original data were incorrectly coded as "D". This has been corrected. Additionally, in 1997, 1998, and 2016 we updated cases of FAMIDRD values of "X" to "B" for greater consistency with other years of data; the previous assignment of these years was still appropriate for unique identification, but the code has been modified.
- We fixed an error that incorrectly reported responding unit identifiers in the 2012 data for the variable RULETRRD: responding unit identifier values that were "A" in previous versions of the data are now "C"; cases that previously had values of "I" are now "A"; cases that previously had values of "C," "D," and "E" were incorrect by one letter and are now assigned values of "D," "E," and "F" respectively. No values were collapsed, so all responding units were still uniquely identifiable using these codes. Additionally, in 1997 and 1998 we updated cases of RULETRRD values of "X" to "B" for greater consistency with other years of data; the previous assignment of these years was still appropriate for unique identification, but the code has been modified.
New data.
Modified data.
Corrected data.
May 20, 2019
- The codes and value labels for a number of variables using a likert scale of "None of the time" to "Most of the time" in the self-administered questionnaire were modified for consistency throughout the entire questionnaire section. "None of the time" is consistently assigned a value of "0", while "All of the time" is now consistently assigned a value of "4". This change affets the following variables: ADDOWN, ADPALS, ADPWLM, ADMALS, ADMWLM, ADCAPE, ADNRGY, and ADSOCA. The three point likert scale for the variables ADDAYA and ADCLIM was also modified to reflect this general pattern, though their scale ranges from "Not limited" to "Limited a lot". Out of universe cases for these variables are assigned to the code of "6."
- The value labels for the variable ADPAIN were incorrect. ADPAIN reports how often pain interefered with normal work in the past 4 weeks and reports values on a likert scale of "None of the time" to "Most of the time"; the value labels of the likert scale were incorrectly inverted. The codes and labels have been updated to reflect the correct direction of the likert scale, and to be consistent with other self-administered questionnaire items using this likert scale.
Modified data.
Corrected data.
April 20, 2019
- The top-codes for WAGEHRRD and WAGENEWRD have been corrected. WAGEHRRD previously assigned all top-coded values to the 1997 top-code value of $43.75, and WAGENEWRD previously assigned all top-coded values to the 2015 top-code value of $80.00. Data now report the appropriate year-specific top code for both variables.
Corrected data.
November 27, 2018
- We have inserted a leading zero in the values of MEPSID and MEPSIDRD in the 1996-2004 samples to replace a leading blank that was preventing users from merging data using formatted data extracts to ASCII data extracts read in using a stats package syntax file. The merging issue previously observed has been resolved.
- The variable PERNUM is now included on round-level records to support hierarchical IPUMS extracts in R.
Corrected data.
New data.
October 4, 2018
- Corrected duplicate cases for 1996 sample.
Modified data.
August 31, 2018
- IPUMS MEPS now includes 2016 FYC data; over 1,300 variables (both annual person-level variables and round-level variables) are now available through the extract system.
- To account for the release of pooled variance estimation variables (STRATAPLD and PSUPLD); the variable names for STRATAP and PSUP were updated to STRATANN and PSUANN to more clearly note that they are designed to produce annual estimates.
- A missing data code for HPTOTNIGHT was updated for consistency with other IPUMS missing data codes; cases previously coded as -9 are now classified as "998: Not ascertained".
- Missing and top-code values for WAGENEWRD were updated for consistency with other IPUMS missing data codes (they previously matched AHRQ missing and top-coded values and were not labeled in data extracts). Cases previously coded as -1.00 are now classified as "0.00: Not in Universe"; cases previously coded as -10.00 are now classified as the respective top-code for each year of data; cases previously coded as -13.00 are now classified as "99.91: Initial wage imputed"; and cases previously coded as -8.00 are now classified as "99.98: Not ascertained".
- The variable WAGEHRRD was expanded by one column and its missing data and top-code values were updated for consistency with other IPUMS missing data codes (they previously matched AHRQ missing data codes and were not-labeled in extracts). Cases previously coded as -1.00 or -3.00 are now classified as "0.00: Not in Universe"; cases previously coded as -2.00 are now classified "9999.94: Determined in previous round"; cases coded as -10.00 are now classified as the respective top-code for each year; cases previously coded as -7.00 are now classified as "9999.97: Refused"; cases previously coded as -8.00 are now classified as "9999.99: Don't know"; and cases previously coded as -9.00 are now classified as "9999.98: Not ascertained".
- The missing data codes for HOURWRKRD in 2001, 2004, and 2007-2009 reflected AHRQ-assigned values, and were not consistent with other IPUMS missing data codes. Cases previously coded as -1 or -3 are now classified as "996: Not in Universe"; cases previously coded as -2 are now classified "994: Determined in previous round"; cases previously coded as -7 are now classified as "997: Refused"; cases previously coded as -8 are now classified as "999: Don't know", and cases previously coded as -8 are now classified as "998: Not ascertained".
- The code for "determined in previous round" was changed from 96 to 94 for the round-level variable WAGEHRSRCRD for greater comparability across variables. The variable is otherwise unchanged.
New data.
Modified data.
August 13, 2018
- Newly released 2015 MEPS data and 300+ additional variables, including round-level, are now integrated into the IPUMS MEPS database.
- EDUCATN has been renamed EDUCYR to match the original MEPS variable name and additional samples were added.
New data.
Modified data.
June 21, 2018
- IPUMS MEPS has updated USBORN and YRSINUS (now named YRSINUSG in 2013 to include both panels 17 and 18. Previously, panel 17 members had been classified as "inapplicable" because of a change in how MEPS offers these variables.
- IPUMS MEPS has corrected an error affecting CPSFAMID. In previous versions of the data, CPSFAMID used incorrect original data for 2008 and 2012, causing the data to be missing in 2012 and incorrect in 2008.
Updated data.
Corrected data.
February 5, 2018
Corrected data.
January 30, 2018
- Beta version of IPUMS MEPS released.
New data.