Documentation for file prov_union
Geographic Spillover of Unionism, January, 2006
Thomas J. Holmes
This file matches records from the 1991, 1996, 2001 POS file to union records from the following sources: FMCS contract data, NLRB election data, contract data from the Department of Labor, information found on the web from union locals making a listing of facilities under contract. The union records are indexed by idunion, a unique identifier for each union record. The POS records are indexed by prov_num year_prov, where prov_num is the unique identifier in the POS file and year_prov is the year, 1991, 1996, 2001.
A variable catlink in the file specifies the category of the union record. The categories are:
|
Catlink |
Type of Link |
|
F- |
fmcs link |
|
N- |
NLRB link |
|
W- |
web link |
|
B- |
from BLS collective
bargaining |
|
s1 |
prov_num is linked to a
prov_num through a longitudinal link |
|
s2 |
prov_num (nhome) is linked
to a prov_num (hospital) through a prov_based_facility=’Y’ from cms. So a nhome that is part of a hospital. |
|
s3 |
address link. There
are cases with the same address, different prov_num. |
|
00 |
=’00’ prov_num has no union links at all |
This is a many to many match. So the same union record can be matched to multiple facilities because a given contract may apply to multiple facilities over time. Furthermore, a given prov_num may be in three times because there are three years of data.
Data Dictionary prov_union
|
FACILITY_NAME |
Name from cms |
|
PROV_NUM |
file number from cms |
|
NUM_BEDS |
from cms |
|
CITY |
from cms. City of
facility |
|
STTEXT |
from cms, state of facility |
|
year_prov |
year from cms |
|
type |
either nhome or hosp |
|
idfmcs |
unique id in FMCS.
Only applies for fmcs records |
|
multi |
is this a union record that applied to multiple
establishments. Left blank most of the
time. |
|
outsource |
=1 if known to be some
outsourcing company. This applied only
to the ones I did by hand. The
automated matches wouln’t have picked this up. |
|
catlink |
See above |
|
employer |
The name of the employer
from the union record |
|
local |
information about the local
from the union record |
|
unionstreet |
|
|
unioncity |
|
|
unionst |
|
|
unionzip |
|
|
unionrep |
|
|
unionphone |
|
|
sizebargain |
for FMCS records, this is
just from FMCS NLRB=number of eligible
voters. |
|
sizeestab |
for FMCS, the sizeestab variable. blank for other records |
|
uniontitle |
title of contract from FMCS |
|
AFF_SHORT |
Ancronym. This is the main union variable that has
been standardized across all the different sources of the data. I merge this variable in wht the parent
information from LORS. |
|
year_union |
year of the union record. For FMCS, I take the following precedence. If 1978<=year_expire <=2007, I use
this. Otherwise I look at notice_emp
variable, otherwise look at notice_uni variable, otherwise look at reopendate
variable |
|
Date_FMCS_flag |
This shows which of the
above is used to create the year_union and Date_FMCS. |
|
Date_FMCS |
Actual date. |
|
id_union |
I unique identifier for
each union record |
|
union |
The name of the union from
the raw data. Use aff_short as the
cleaned up union name. |
|
idweb |
unique identified from the
web data |
|
idnlrb |
link in the NLRB file |
|
union_index |
A special variable created
for the 77-99 NLRB recrods. Some
recrrods have muliple unions. I creaed
a separate record for each union (to be like 00-02 records). This is the index for each uino of each multpile
union election. |
|
Barg_Unit |
Kind of unit:
proffessional, truack drivers, etc. See CATS documentation for
the codes |
|
Unit_id |
NLRB 00-02 variable (for
elections with the same case nmber, but multiple elections at the same time
for different bargaining units).
Created an analogous variable for 77-99 |
|
Elect_Result |
WON or LOST |
|
Incumbent |
Indicator variable =1 if
incumbent. Not a clean variable for
2000-2003. |
|
Role |
P (petitioner) or I
(intervenor). Clean from 77-99. For 2000-2003 had to pull in R_participant
file. |
|
R_Case_Number |
NLRB number |
|
Num_Valid_Votes |
|
|
VotesAgainst |
vote for no union |
|
Case_Type |
RC, RD, RM |
|
Date_Election |
|
|
Votes_For |
votes for the union |
|
RptUnions |
Number of unions in the
election (a 00-02 variable that I created for 77-99) |
|
union_code_nlrb7799 |
the union code used in the
77-99 file |
|
PROV_NUM_LINK |
if the union record goes with another prov_num (s1,
s2,s3) this is what the other one is. |
|
year_prov_link |
This is the year of the
other prov_num link |
|
AFF_CODE |
this is the aff_code of the aff_short in the parent. |
|
AFF_NAME |
this is the aff_name that
goes with the aff_short |
|
AFF_FN |
this is the aff_fn that I crated
in in the LORS parent |
|
F_NUMBER |
this if the f_number that goes with those records for
which I found matches but were not aff_names.
So suppose LORS treats a union as an independent. Then there is no aff_name. But I found some of these LORS records for
these guys, so this is the f_number that matched up. |