Skip navigation to content
eriu: Economic Research Initiative on the Uninsured Initiating and disemminating research to spark new policy discussion on health coverage issues.
Research Findings  
Fast Facts Home
Facts to Consider
Counting the Uninsured by Reference Period
All Year
Ever in Year
Point in Time
By Data Source
By Population
Total Non-Elderly
By Additional Characteristics

Education Level
Family Composition

Comparison Tables
Medical Utilization and
Expenditures Tables
Notes on Data Sources
and Variables
Data Dictionary

Notes on MEPS Data Source and Variables

Source data: 2000 and 2001 Full Year Population Characteristics of the Medical Expenditure Panel Survey. All data are weighted by the final person weight, PERWT00F.

Gender is determined using the variable SEX.

Age is determined using the variable AGE31X.

Race and Hispanic Status are determined using the variables RACEX and HISPANX. Some of the tables include Hispanic as a separate “race,” others incorporate it.

Income as percentage of federal poverty level is determined using POVCAT00. Categories reported for MEPS differ slightly from categories reported for CPS and SIPP because the category 200-300% is not included in the POVCAT00 variable.

Education is determined using the variables EDUCYEAR and HIDEGYR.

Wage is determined using HRWG31X, which is topcoded at the hourly wage of $57.50.

Insurance status is determined using several variables. Full year uninsured is defined using the value of 3 for INSCOV00. Ever uninsured during the year was created using INSJA00X through INSDE00X, monthly insurance indicators. Point in time status was determined by INS31X, an indicator of whether the person was uninsured during the interview round (each rounds spans about 3 to 6 months). The point in time used to determine point in time insurance status, age, education, work status and family composition is the first interview round during the year.

Family Composition is determined using the variables RFREL31X and SPOUID31. RFREL31X indicates the relationship to the reference person. An RFREL31X value of 0 indicates the person is the respondent. For each reporting unit, the person who owns or rents the dwelling unit is usually defined as the reference person. SPOUID31, specifically value 995- no spouse in house, is used to identify married couples where the spouse is not present. Family is identified by the variable CPSFAMID, which uses the CPS definition of family. This was so that estimates would be comparable to SIPP and CPS estimates, and because the POVCAT00 variable is defined based on cps family definition.

Family work status is determined using HOUR31, EMPST31 and SELFCM. A person is part time if they work between 1 and 35 hours per week, and full time if they work 35 hours or more per week. Those with an EMPST31 value of 3, who did not have a job at the time of interview but did have a job during the round, were coded as unemployed because they did not have data on hours worked or wage. Point in time status is determined based on status at the first interview round during the year.