Extract Data Structures Available through IPUMS MEPS

Rectangular on Person (Default)

• There is one record per person, per year.


• Only person-level variables (i.e., those listed under the "Annual" drop down menu) can be included in this type of extract. Event-level and condition-level variables cannot be included in this type of extract.

Figure 1: Example extract formatted as rectangular on person

mepsid	panel	year	relyr	age	povcat	smokenow
2070001101	20	2015	1	65	4	1
2070001102	20	2015	1	57	4	1
2070004102	20	2015	1	40	5	1
2070005101	20	2015	1	55	2	2
2070005102	20	2015	1	21	2	0
2070006101	20	2015	1	26	5	0
2070006103	20	2015	1	53	5	0
2070006104	20	2015	1	57	5	1
2070010101	20	2015	1	41	4	1

Rectangular on Round

• There is one record per round, person, and year. In a given year of data, persons have three round-level records.


• Both person-level and round-level variables can be included in this type of extract. Event-level and condition-level variables cannot be included in this type of extract.

Figure 2: Example extract formatted as rectangular on round

mepsid	panel	year	relyr	age	povcat	smokenow	roundrd	agerd	empstatrd
2070001101	20	2015	1	65	4	1	1	64	0
2070001101	20	2015	1	65	4	1	2	65	0
2070001101	20	2015	1	65	4	1	3	65	1
2070001102	20	2015	1	57	4	1	1	57	4
2070001102	20	2015	1	57	4	1	2	57	4
2070001102	20	2015	1	57	4	1	3	57	4
2070005101	20	2015	1	40	5	1	1	39	1
2070005101	20	2015	1	40	5	1	2	40	1
2070005101	20	2015	1	40	5	1	3	40	1

Back to Top

Hierarchical

• This extract type can contain person-level, round-level, condition-level, and event-level records. Per year, there is a single person-level record containing only person-level variables (those listed under the "Annual" drop down menu) followed by three round-level records containing only round-level variables (those listed under the "Round" drop down menu).


• Person-level, round-level, condition-level, and event-level variables can be included in this type of extract. This is the only extract type that can include condition-level and event-level variables.

Figure 3. Example extract that is formatted as hierarchical

rectype	mepsid	panel	year	age	roundrd	agerd	mri	icd9code	eventypej
P	2070001101	20	2015	65	.	.	.	.	.
R	.	.	.	.	1	64	.	.	.
R	.	.	.	.	2	65	.	.	.
R	.	.	.	.	3	65	.	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No	.	.
E	.	.	.	.	.	.	No	.	.
E	.	.	.	.	.	.	NIU	.	.
E	.	.	.	.	.	.	NIU	.	.
E	.	.	.	.	.	.	No services received	.	.
X	.	.	.	.	.	.	.	272	.
J	.	.	.	.	.	.	.	.	8
J	.	.	.	.	.	.	.	.	8
X	.	.	.	.	.	.	.	9999	.
J	.	.	.	.	.	.	.	.	1
X	.	.	.	.	.	.	.	No services received	.
J	.	.	.	.	.	.	.	259	.
J	.	.	.	.	.	.	.	.	8
J	.	.	.	.	.	.	.	.	8
J	.	.	.	.	.	.	.	.	8
X	.	.	.	.	.	.	.	401	.
P	2070001102	20	2015	57	.	.	.	.	.
R	.	.	.	.	1	57	.	.	.
R	.	.	.	.	2	57	.	.	.
R	.	.	.	.	3	57	.	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No services received	.	.
E	.	.	.	.	.	.	No	.	.
E	.	.	.	.	.	.	NIU	.	.
X	.	.	.	.	.	.	.	272	.
J	.	.	.	.	.	.	.	.	8
J	.	.	.	.	.	.	.	.	8
X	.	.	.	.	.	.	.	799	.
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	8
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	1
J	.	.	.	.	.	.	.	.	8
X	.	.	.	.	.	.	.	V77	.
J	.	.	.	.	.	.	.	.	8
J	.	.	.	.	.	.	.	.	8

Wide on Person

• There is one record per person, per year.


• This extract type contains both person and round-level variables, written to person-level records. Event-level and condition-level variables cannot be included in this type of extract.


• The names of variables in this data format type differ from the variable names received with other data formats. The variable names include numeric suffixes indicating the time period that the information covers. For example, on variables that were originally person level, an _1 or _2 will appear at the end of the variable name, referring to the first and second year of MEPS data collection, respectively (see RELYR). On variables that were originally round level, a set of numeric suffixes ranging from _1_1 to _2_5 are appended to each variable name. The first number refers to the first or second year of MEPS data collection and the second number refers to the round of data collection. For example, _1_1 refers to variables measured during the first year of data collection and the Round 1 interview, and _2_5 refers to variables measured during the second year of data collection and the Round 5 interview.

Figure 4. Example extract that is formatted as wide on person

mepsid	year	age_1	age_2	agerd_1_1	agerd_1_2	agerd_1_3	agerd_2_3	agerd_2_4	agerd_2_5
2070001101	2015	65	.	64	65	65	.	.	.
2070001101	2016	.	66	.	.	.	65	66	66
2070001102	2015	57	.	57	57	57	.	.	.
2070001102	2016	.	58	.	.	.	57	58	58
2070004102	2015	40	.	39	40	40	.	.	.
2070004102	2016	.	41	.	.	.	40	41	41

Back to Top