= LEHD Public Use Data Schema documents
Lars Vilhuber <lars.vilhuber@cornell.edu>
28 February 2019
// a2x: --dblatex-opts "-P latex.output.revhistory=0 --param toc.section.depth=3"
:ext-relative: {outfilesuffix}
( link:lehd_public_use_schema.pdf[Printable version] )



[IMPORTANT]
.Important
==============================================
This document is not an official Census Bureau publication. It is compiled from publicly accessible information
by Lars Vilhuber (http://www.ilr.cornell.edu/ldi/[Labor Dynamics Institute, Cornell University]).
Feedback is welcome. Please write us at
link:mailto:lars.vilhuber@cornell.edu?subject=LEHD_Schema_v4[lars.vilhuber@cornell.edu].
==============================================
	

Purpose
-------
The public-use data from the Longitudinal Employer-Household Dynamics Program, including the Quarterly Workforce Indicators (QWI)
and Job-to-Job Flows (J2J), are available for download with the following data schema.
These data are available  through the LEHD website’s Data page at
http://lehd.ces.census.gov/data/ and through the LED Extraction Tool at http://ledextract.ces.census.gov/.

This document describes the data schema for LEHD files. LEHD-provided SHP files are separately described in link:lehd_shapefiles{ext-relative}[]. For each variable,
a set of allowable values is defined. Definitions are provided as CSV files,
with header variable definitions.  Changes relative to the original v4.0 version are listed <<changes,at the end>>.

File naming
-----------
The naming conventions of the data files is documented in link:lehd_csv_naming{ext-relative}[].

Extends
-------
This version reimplements some features from  V4.0. Many files compliant with LEHD or QWI Schema v4.0 will also be compliant with this schema, but compatibility is not guaranteed.

Supersedes
----------
This version supersedes V4.2.0, for files released as of R2018Q3.

Basic Schema
------------
Each data file is structured as a CSV file. The first columns contain <<identifiers>>, subsequent columns contain <<indicators>>, followed by <<statusflags,status flags>>. In some cases, visually formatted Excel (XLSX) files are also available,  containing the same information together with header lines  on each sheet.

=== Generic structure

[width="30%",format="csv",cols="<2",options="header"]
|===================================================
Column name
[ Identifier1 ]
[ Identifier2 ]
[ Identifier3 ]
[ ... ]
[ Indicator 1 ]
[ Indicator 2 ]
[ Indicator 3 ]
[ ... ]
[ Status Flag 1 ]
[ Status Flag 2 ]
[ Status Flag 3 ]
[ ... ]
|===================================================

Note: A full list of indicators for each type of file are shown below in the <<indicators,Indicators>> section.
While all indicators are included in the CSV files, only the requested indicators
will be included in data outputs from the LED Extraction Tool.

<<<

=== [[identifiers]]Identifiers
Records, unless otherwise noted, are parts of time-series data. Unique record identifiers are noted below, by file type.
Identifiers without the year and quarter component can be considered a series identifier.

==== Mapping for Identifiers
( link:lehd_mapping_identifiers.csv[] )

Each of the released files has a set of variables uniquely identifying records ('Identifiers'). The table below relates the set of identifier specifications
to the released files. The actual CSV files containing the identifiers for each set are listed after this table. Each identifier can take on a specified list of values, documented in the section on <<catvars,Categorical Variables>>.

[width="80%",format="csv",cols="<3,6*^1",options="header"]
|===================================================
include::lehd_mapping_identifiers.csv[]
|===================================================
<<<

==== Identifiers for j2j
( link:lehd_identifiers_j2j.csv[] )

[width="100%",format="csv",cols="2*^1,<3",options="header"]
|===================================================
include::lehd_identifiers_j2j.csv[]
|===================================================
<<<


==== Identifiers for j2jod
( link:lehd_identifiers_j2jod.csv[] )

[width="100%",format="csv",cols="2*^1,<3",options="header"]
|===================================================
include::lehd_identifiers_j2jod.csv[]
|===================================================
<<<


==== Identifiers for qwi
( link:lehd_identifiers_qwi.csv[] )

[width="100%",format="csv",cols="2*^1,<3",options="header"]
|===================================================
include::lehd_identifiers_qwi.csv[]
|===================================================
<<<



<<<
=== [[indicators]]Indicators
The following tables and associated mapping files
list the indicators available on each file.  The descriptor files themselves are structured as follows:

- The ''Indicator Variable'' is the short name of the variable on the CSV files, suitable for machine processing in a wide variety of statistical applications.
- When given, the ''Alternate name'' may appear in related documentation and articles.
- The ''Status Flag'' is used to indicate publication or data quality status (see <<statusflags,Status Flags>>).
- The ''Indicator Name'' is a non-abbreviated version of the ''Indicator Variable''.
- The ''Description'' provides more verbose description of the variable.
- ''Units'' identify the type of variable according to a very simplified taxonomoy (not formalized yet): counts, rates, monetary amounts.
- ''Concept'' classifies the variables into higher-level concepts. The taxonomy for these concepts has not been finalized yet, see link:label_concept_draft.csv[label_concept_draft.csv] for a draft version.
- The ''Base'' indicates the denominator used to compute the statistic, and may be '1'.

==== National QWI and state-level QWI (QWIPU) ====

( link:variables_qwi.csv[variables_qwi.csv] )
[width="95%",format="csv",cols="3*^2,<5,<5,<2,<2,^2",options="header"]
|===================================================
include::variables_qwi.csv[]
|===================================================
<<<

==== National QWI and state-level QWI rates (QWIPUR) ====
Rates are computed from published data, and are provided as a convenience.


( link:variables_qwir.csv[variables_qwir.csv] )
[width="95%",format="csv",cols="3*^2,<5,<5,<2,<2,<2",options="header"]
|===================================================
include::variables_qwir.csv[]
|===================================================


<<<

==== Job-to-job flow counts (J2J)
( link:variables_j2j.csv[] )
[width="95%",format="csv",cols="3*^2,<5,<5,<2,<2,^1",options="header"]
|===================================================
include::variables_j2j.csv[]
|===================================================
<<<

==== Job-to-job flow rates (J2JR)
( link:variables_j2jr.csv[] )

Rates are computed from published data, and are provided as a convenience.


[width="95%",format="csv",cols="3*^2,<5,<5,<2,<2,^1",options="header"]
|===================================================
include::variables_j2jr.csv[]
|===================================================


<<<

==== Job-to-job flow Origin-Destination (J2JOD)
( link:variables_j2jod.csv[] )
[width="95%",format="csv",cols="3*^2,<5,<5,<2,<2,^1",options="header"]
|===================================================
include::variables_j2jod.csv[]
|===================================================
<<<

<<<


<<<
=== [[vmeasures]]Variability measures
The following tables and associated mapping files
list the variability measures available on each file.  The ''Variability Measure'' is the short name of the variable on the CSV files,
suitable for machine processing in a wide variety of statistical applications. When given, the ''Alternate Name'' may appear in related documentation and articles.
The ''Variable Name'' is a more verbose description of the variability measure.

Six variability measures are published:

* Total variability, prefixed by vt_
* Standard error, prefixed by st_, and computed as the square root of Total Variability
* Between-implicate variability, prefixed by vb_
* Average within-implicate variability, prefixed by vw_
* Degrees of freedom, prefixed by df_
* Missingness ratio, prefixed by mr_

A missing variability measure indicates a structural zero in the corresponding indicator. This is currently not associated with a flag.

//Not all indicators have associated variability measures. For more details, see the following document TBD.

==== Generic structure

[width="30%",format="csv",cols="<2",options="header"]
|===================================================
Column name
[ Identifier1 ]
[ Identifier2 ]
[ Identifier3 ]
[ ... ]
[ Standard error for Indicator 1 ]
[ Standard error for Indicator 2 ]
[ Standard error for Indicator 3 ]
[ ... ]
[ Total variation for Indicator 1 ]
[ Total variation for Indicator 2 ]
[ Total variation for Indicator 3 ]
[ ... ]
[ Between-implicate variability for Indicator 1 ]
[ Between-implicate variability for Indicator 2 ]
[ Between-implicate variability for Indicator 3 ]
[ ... ]
[ Average within-implicate variability for Indicator 1 ]
[ Average within-implicate variability for Indicator 2 ]
[ Average within-implicate variability for Indicator 3 ]
[ ... ]
[ Degrees of freedom for Indicator 1 ]
[ Degrees of freedom for Indicator 2 ]
[ Degrees of freedom for Indicator 3 ]
[ ... ]
[ Missingness ratio for Indicator 1 ]
[ Missingness ratio for Indicator 2 ]
[ Missingness ratio for Indicator 3 ]
[ ... ]
|===================================================


Note: A full list of indicators for each type of file are shown  in the <<indicators,Indicators>> section. In the tables below, only a sample
of variability measures are printed, but the complete list is available in the linked CSV schema files.

<<<

==== National QWI and state-level QWI ====

( link:variables_qwiv.csv[variables_qwiv.csv] )
[width="95%",format="csv",cols="2*^2,<5,<5,<2",options="header"]
|===================================================
include::tmp_variables_qwiv.csv[]
|===================================================

<<<
==== National QWI and state-level QWI rates ====

( link:variables_qwirv.csv[variables_qwirv.csv] )
[width="95%",format="csv",cols="2*^2,<5,<5,<2",options="header"]
|===================================================
include::tmp_variables_qwirv.csv[]
|===================================================


<<<

==== Job-to-job flow counts (J2J)
Soon.
//( link:variables_j2j.csv[] )
//[width="95%",format="csv",cols="3*^2,<5",options="header"]
//|===================================================
//include::tmp_variables_j2jv.csv[]
//|===================================================
//<<<
//

==== Job-to-job flow rates (J2JR)
Soon.
//( link:variables_j2jr.csv[] )
//[width="95%",format="csv",cols="3*^2,<5",options="header"]
//|===================================================
//include::tmp_variables_j2jrv.csv[]
//|===================================================
//<<<

==== Job-to-job flow Origin-Destination (J2JOD)
Soon.
//( link:variables_j2jod.csv[] )
//[width="95%",format="csv",cols="^3,^2,^3,<5",options="header"]
//|===================================================
//include::tmp_variables_j2jodv.csv[]
//|===================================================

<<<



== [[catvars]]Categorical Variables
Categorical variable descriptions are displayed above each table, with the variable name shown in parentheses. Unless otherwise stated, every possible value/label combination for each categorical variable is listed. Please note that not all values will be available in every table.


=== agegrp
( link:label_agegrp.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_agegrp.csv[]
|===================================================

=== education
( link:label_education.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_education.csv[]
|===================================================

=== ethnicity
( link:label_ethnicity.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_ethnicity.csv[]
|===================================================

=== firmage
( link:label_firmage.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_firmage.csv[]
|===================================================

=== firmsize
( link:label_firmsize.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_firmsize.csv[]
|===================================================

=== ownercode
( link:label_ownercode.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_ownercode.csv[]
|===================================================

=== periodicity
( link:label_periodicity.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_periodicity.csv[]
|===================================================

=== quarter
( link:label_quarter.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_quarter.csv[]
|===================================================

=== race
( link:label_race.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_race.csv[]
|===================================================

=== seasonadj
( link:label_seasonadj.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_seasonadj.csv[]
|===================================================

=== sex
( link:label_sex.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_sex.csv[]
|===================================================

<<<
=== Industry ===

 
[[ind_level]]
==== Industry levels
( link:label_ind_level.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_ind_level.csv[]
|===================================================


==== Industry
( link:label_industry.csv[] )

Only a small subset of available values shown.
The 2017 NAICS (North American Industry Classification System) is used for all years.
QWI releases prior to R2018Q1 used the 2012 NAICS classification (see link:../V4.1.3[Schema v4.1.3]).
For a full listing of all valid 2017 NAICS codes, see http://www.census.gov/cgi-bin/sssd/naics/naicsrch?chart=2017.

[width="90%",format="csv",cols="^1,<5,^1",options="header"]
|===================================================
include::tmp2.csv[]
|===================================================
<<<

=== [[geography]]Geography ===

  
[[geo_level]]
==== [[geolevel]] Geographic levels
Geography labels for data files are provided in separate files, by scope. Each file 'label_geograpy_SCOPE.csv' may contain one or more types of records as flagged by <<geolevel,geo_level>>. For convenience, a composite file containing all geocodes is available as link:label_geography.csv[].
The 2017 vintage of  https://www.census.gov/geo/maps-data/data/tiger-line.html[Census TIGER/Line geography] is used for all tabulations as of the R2018Q1 release.


Shapefiles are described in a link:lehd_shapefiles{ext-relative}[separate document].


( link:label_geo_level.csv[] )

[width="80%",format="csv",cols="^1,<3,<8,<8",options="header"]
|===================================================
include::label_geo_level.csv[]
|===================================================



==== [[geostate]]National and state-level values ====
( link:label_fipsnum.csv[] )

The file link:label_fipsnum.csv[label_fipsnum.csv] contains values and labels
for all entities of <<geolevel,geo_level>> 'N' or 'S', and is a summary of separately available files.

[width="40%",format="csv",cols="^1,<3,^1",options="header"]
|===================================================
include::tmp.csv[]
|===================================================

==== [[stusps]]State postal codes

Some parts of the schema use (lower or upper-case) state postal codes.

( link:label_stusps.csv[] )

[width="60%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_stusps.csv[]
|===================================================


==== [[geosubstate]]Detailed state and substate level values

Files of type 'label_geography_[ST].csv' will contain identifiers and labels for geographic areas entirely comprised within a given state '[ST]'. State-specific parts of cross-state CBSA, in records of type <<geolevel,geo_level>> = M, are present on files of type 'label_geography_[ST].csv'. The file link:label_geography_metro.csv[] contains labels for records of type <<geolevel,geo_level>> = B, for metropolitan areas only.





[format="csv",width="50%",cols="^1,^2,^3",options="header"]
|===================================================
Scope,Types,Format file
US,N,link:label_geography_us.csv[]
METRO,B,link:label_geography_metro.csv[]
*States*,,
AK,S C W M,link:label_geography_ak.csv[]
AL,S C W M,link:label_geography_al.csv[]
AR,S C W M,link:label_geography_ar.csv[]
AZ,S C W M,link:label_geography_az.csv[]
CA,S C W M,link:label_geography_ca.csv[]
CO,S C W M,link:label_geography_co.csv[]
CT,S C W M,link:label_geography_ct.csv[]
DC,S C W M,link:label_geography_dc.csv[]
DE,S C W M,link:label_geography_de.csv[]
FL,S C W M,link:label_geography_fl.csv[]
GA,S C W M,link:label_geography_ga.csv[]
HI,S C W M,link:label_geography_hi.csv[]
IA,S C W M,link:label_geography_ia.csv[]
ID,S C W M,link:label_geography_id.csv[]
IL,S C W M,link:label_geography_il.csv[]
IN,S C W M,link:label_geography_in.csv[]
KS,S C W M,link:label_geography_ks.csv[]
KY,S C W M,link:label_geography_ky.csv[]
LA,S C W M,link:label_geography_la.csv[]
MA,S C W M,link:label_geography_ma.csv[]
MD,S C W M,link:label_geography_md.csv[]
ME,S C W M,link:label_geography_me.csv[]
MI,S C W M,link:label_geography_mi.csv[]
MN,S C W M,link:label_geography_mn.csv[]
MO,S C W M,link:label_geography_mo.csv[]
MS,S C W M,link:label_geography_ms.csv[]
MT,S C W M,link:label_geography_mt.csv[]
NC,S C W M,link:label_geography_nc.csv[]
ND,S C W M,link:label_geography_nd.csv[]
NE,S C W M,link:label_geography_ne.csv[]
NH,S C W M,link:label_geography_nh.csv[]
NJ,S C W M,link:label_geography_nj.csv[]
NM,S C W M,link:label_geography_nm.csv[]
NV,S C W M,link:label_geography_nv.csv[]
NY,S C W M,link:label_geography_ny.csv[]
OH,S C W M,link:label_geography_oh.csv[]
OK,S C W M,link:label_geography_ok.csv[]
OR,S C W M,link:label_geography_or.csv[]
PA,S C W M,link:label_geography_pa.csv[]
RI,S C W M,link:label_geography_ri.csv[]
SC,S C W M,link:label_geography_sc.csv[]
SD,S C W M,link:label_geography_sd.csv[]
TN,S C W M,link:label_geography_tn.csv[]
TX,S C W M,link:label_geography_tx.csv[]
UT,S C W M,link:label_geography_ut.csv[]
VA,S C W M,link:label_geography_va.csv[]
VT,S C W M,link:label_geography_vt.csv[]
WA,S C W M,link:label_geography_wa.csv[]
WI,S C W M,link:label_geography_wi.csv[]
WV,S C W M,link:label_geography_wv.csv[]
WY,S C W M,link:label_geography_wy.csv[]
|===================================================

<<<
=== Aggregation level
( link:label_agg_level.csv[] )

Measures within the J2J and QWI data products are tabulated on many different dimensions, including demographic characteristics, geography, industry, and other firm characteristics. For Origin-Destination (O-D) tables, characteristics of the origin and destination firm can be tabulated separately.  Every tabulation level is assigned a unique aggregation index, represented by the agg_level variable. This index starts from 1, representing a national level grand total (all industries, workers, etc.), and progresses through different combinations of characteristics. There are gaps in the progression to leave space for aggregation levels that may be included in future data releases.

*agg_level* is currently  reported only for  J2J data products.


The following variables are included in the link:label_agg_level.csv[label_agg_level.csv]   file:

[width="60%",format="csv",cols="<2,<5",options="header"]
|===================================================
include::variables_agg_level.csv[]
|===================================================


The characteristics available on an aggregation level are repeated using a series of flags following the standard schema:

- <<geolevel,geo_level>> - geographic level of table
- <<ind_level,ind_level>> - industry level of table
- by_ variables - flags indicating other dimensions reported, including ownership, demographics, firm age and size.

A shortened representation of the file is provided below, the complete file is available in the link above.


[width="90%",format="csv",cols=">1,3*<2,5*<1",options="header"]
|===================================================
include::tmp_label_agg_level.csv[]
|===================================================


<<<
== [[statusflags]]Status Flags
( link:label_flags.csv[] )

Each status flag in the tables above contains one of the following valid values.
The values and their interpretation are listed in the table below.

[IMPORTANT]
.Important
==============================================
Note: Currently, the J2J tables only contain status flags '-1', '1', '5'. Status flags with values 10 or above only appear in online applications, not in CSV files.
==============================================


[width="80%",format="csv",cols="^1,<4",options="header"]
|===================================================
include::label_flags.csv[]
|===================================================


<<<

== [[metadata]]Metadata
( link:variables_version.csv[] )

=== [[metadataqwij2j]]Version Metadata for QWI and J2J Files (version.txt)

Each data release is accompanied by one or more files with metadata on geographic and temporal coverage, in a compact notation. These files follow the following naming convention:
--------------------------------
version_[type].txt
--------------------------------
where each component is described in more detail in link:lehd_csv_naming{ext-relative}[].

The contents contains the following elements:
[width="90%",format="csv",cols="<1,<2,<5",options="header"]
|===================================================
include::tmp_variables_version.csv[]
|===================================================

For instance, the metadata for the R2018Q3 QWI release of
Delaware
(obtained from https://lehd.ces.census.gov/data/qwi/R2018Q3/de/version_qwi.txt[here]) has  the following content:
--------------------------------


QWI_F DE 10 1998:3-2017:4 V4.3.0 R2018Q3 qwipu_de_20180813_1413
QWI_FA DE 10 1998:3-2017:3 V4.3.0 R2018Q3 qwipu_de_20180813_1413
QWI_FS DE 10 1998:3-2017:3 V4.3.0 R2018Q3 qwipu_de_20180813_1413
--------------------------------
Similarly, the metadata for the R2018Q2 release of
Delaware J2J
tabulations (obtained from https://lehd.ces.census.gov/data/j2j/R2018Q2/j2j/de/version_j2j.txt[here]) has  the following content:
--------------------------------
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head>
<title>301 Moved Permanently</title>
</head><body>
<h1>Moved Permanently</h1>
<p>The document has moved <a href="http://lehd.ces.census.gov/data/j2j/R2018Q2/de/j2j/version_j2j.txt">here</a>.</p>
</body></html>
--------------------------------
Some J2J metadata may contain multiple lines, as necessary.

=== [[metadataj2jod]]Additional Metadata for J2JOD Files (avail.csv)
(link:variables_avail.csv[])

Because the origin-destination (J2JOD) data link two regions, we provide an auxiliary file with the time range that cells containing data for each geographic pairing may appear in a data release.
[width="80%",format="csv",cols="<2,<2,<4",options="header"]
|===================================================
include::variables_avail.csv[]
|===================================================
The reference region will always be either the origin or the destination. National tabulations contain records where both origin and destination are <<geolevel,geo_level>>=N; state tabulations contain records where <<geolevel,geo_level>> in (N,S); metro tabulations contain records where <<geolevel,geo_level>> in (N,S,B). Data may be suppressed for certain combinations of regions and quarters because the estimates do not meet Census Bureau publication standards.



=== [[metadatalags]]Metadata on Indicator Availability
(link:variables_lags.csv[])

Each <<indicators,Indicator>> potentially requires leads and/or lags of data to be computed, and thus may not be available for certain time periods. Only two QWI will be available for all quarters of the time span described by +start+ and +end+ in the <<metadataqwij2j,version.txt>> files:  +EmpTotal+ and +Payroll+.  The date range for QWI, QWIR, J2J, and J2JR can be found in <<metadataqwij2j,version.txt>>; the date range for J2JOD can be found in <<metadataj2jod,avail.csv>>.

For each indicator, the following files contain the quarters of data required to be available relative to the overall date range described in the metadata for the release:

* link:lags_qwi.csv[]
* link:lags_j2j.csv[]

The files are structured as follows:
[width="80%",format="csv",cols="<2,<2,<4",options="header"]
|===================================================
include::variables_lags.csv[]
|===================================================
<<<


== [[changes]] Changes
For a description of how schema files are versioned, see link:../VERSIONING{ext-relative}[main directory].

=== This version (revisions)
- 2018-10-30: Initial release
- 2019-02-28: Switch to different rendering engine, only stylistic changes

=== Version 4.3.1 from 4.3.0
- 2018-10-30: Due to a change of the geographic boundaries for a WIB   in Iowa, the shape files need to be updated.
- Adjust the "supersedes" definition to correctly reference the prior version.

=== Version 4.3.0 from 4.2.0
- Changed the format of the version.txt file. This does not affect data files, but it does affect the metadata, and thus requires an increase in the minor version.
- Changed the description of the download locations of the QWI files, as part of the "naming" schema.


<<<
*******************
This revision: Thu Feb 28 21:17:06 UTC 2019
*******************