Newsletter

August 1999 - Number 4

FLS Data Updates

The following data updates have become available since the preceding issue of the FLS Newsletter. For additional updates that may have occurred since the publication of this newsletter, as well as earlier updates you may have missed, please check the respective survey update page on the FLS home page. All updates are dated.

The online data are available via the FLS home page and via RAND's anonymous ftp server (see "Obtaining FLS Data"). All files available online are the most up-to-date versions. The IFLS data on CD-ROM from RAND and the MFLS data on magnetic tape will not be revised; the revised files mentioned in the updates must be obtained from the above online sources. The Inter-University Consortium for Political and Social Research, which also distributes the FLS data, has been notified of updates and has received corrected data.

Stata Versions Available for All FLS Data

For the convenience of FLS users, we have created Stata versions of all public-release databases: MFLS-1, MFLS-2, IFLS-1, EGSF, and MHSS. These versions are available from the RAND ftp site and via the FLS Web page. Please see the "Obtaining FLS Data" section for details.

MFLS-2 Updates

Date: 7/97

In the Panel and Child MF22PREG file for CASE=970,SPLIT=0, the following corrections have been made:

IF EVENTNO=1 THEN CHILD_ID=7;
IF EVENTNO=2 THEN CHILD_ID=6;

Date: 9/98

The SAS transport and ASCII versions of the TRACKING data for the Panel and Child sample have been updated to correct a few MFLS1 identifier codes for children not found on SPLIT=2 records.

Date: 6/99

In the MF22PREG file and the MF22CONT file for both New and Senior and Panel and Children samples, a few cases with duplicate CHILD_ID values were found. Initially, the code to fix those values was listed under MFLS2 updates. The online versions of the MF22PREG and MF22CONT have been corrected and no longer have duplicate CHILD_ID values. The SAS transport, Stata, and ASCII versions of the data have been updated.

MHSS Updates

Date: 8/2/99

The SAS transport and Stata versions of the following MHSS subfiles were corrected for variables inadvertently omitted when the public use files were created. The mhdexp3, mhdexp5, dnfsexp3, dnfsexp5, migexp3, migexp5, mhd3_dta, mhd5_dta, dnf3_dta, dnf5_dat, mig3_dta, and mig5_dta zip files have been updated accordingly.

MHDCM1	MHDOC1	MHDTF1	MHDCCM1	MHDCOC1
MIGCM1	MIGOC1	MIGTF1	MIGCCM1	MIGCOC1
DNFSCM1	DNFSOC1	DNFSTF1	DNFSCCM1	DNFSCOC1

The book3b.doc codebook file was corrected to reorder the names Ch25A and Ch25AU in section CH. Variable Ch25AU is "type of land" and Ch25A is "how much land." In the book3b.doc, book4.doc, and book5.doc codebook files, the codes for ISN302, ISN303, ISN402, ISN403, ISN502, and ISN503 are 1-5, not A-E. The codebook files in hhcdbk.zip have been corrected.

IFLS-1 Updates

Date: 12/96

The new file BA_CHKID added on 9/20/96 had an error. Some children had missing CH05 values. Those values have now been corrected.

In the IFLS-HH codebook, the code definitions for RJ01B (page 256 in Book 3) and RJA01B (page 433 in Book 5) should be as follows to match the IFLS-HH questionnaire:

• 1 = Public Health Center
• 3 = Branch of Public Health Center (not a Puskesmas Pembantu)
• 5 = None

Date: 4/97

The variable ASSETTYP in the BUK2HR1 SAS transport subfile was incorrectly created and has been corrected. The data in the ASCII version have always been correct.

The variable DFACTYPE in the IFLS-CF file BUKPJ01 was in error, coding some Puskesmas as hospitals. The data have been revised.

The variable FACCODE has been added to the IFLS-CF files BUKIJ01 and BUKPJ01. The assignment of FACCODE to nonhospital providers in section J of the Book 1 and Book PKK Community-Facility data was completed by a RAND research project and is being made publicly available. We thank the staff of the research project for their generosity.

Date: 5/97

The variable FACCODE was revised for some records in the BUK1PP1, BUK3RJ2, BUK3RJN, BUK4CX1, BUK4KL4B, and BUK5RJ2 files after a review by a RAND project looking at facility availability.

A new "fixes" file has been added: BUK3MG2.FIXES. It has code to correct cases that transposed MG24MTH to a year value and MG24YR to a month value.

Additional "other,specify" text response records were located and have been appended to the OTHTRANS files. OTHTRANS now has 96,497 records. The OTHTRANS subfiles have been revised.

Date: 1/98

In the IFLS-CF codebook, the code definitions for the months January through December were listed erroneously as 1-12. For variables BE8 and BE9 (pages 65 and 66), the code definitions should be as follows:

A = January	G = July
B = February	H = August
C = March	I = September
D = April	J = October
E = May	K = November
F = June	L = December

The above codes match those in the IFLS-CF questionnaire. The PDF and hardcopy versions have not been updated, so please make the revisions to your hardcopy.

Date: 2/98

An error in character-string wrapping was found in the BUK4KL3.FIX "fixes" file. The problem has been fixed, and the online files have been updated.

Date: 8/98

The following are revisions for the IFLS-CF codebook. The PDF files have not been updated, so please make the revisions to your hardcopy.

For variables JB12A through JB12H, the correct code definitions are

• 1 = Public
• 3 = Private

For variables JB13A to JB13F, the following additional code definitions should be added:

• 4 = morning and afternoon
• 6 = morning and evening
• 8 = afternoon and evening
• 9 = all day

Date: 3/99

A duplicate record found in file BUKKAR3 for CASE=332918622, AR001A=3 was deleted. There are now 33,081 records in BUKKAR3, and the BUKKAR3 files have been updated.

EGSF Updates

Date: 6/99

Note that in PEMITB01, 4 women have B17=1 (speak another language) and B16=B18 (language spoken at home is same as other language spoken). It is not clear whether these 4 women are truly bilingual and the language in either B16 or B18 was misrecorded, or whether these women are truly monolingual and B17 was misrecorded. The EGSF project team does not know where the miscoding occurred and are not revising the data, but are alerting users to the 4 cases. In the codebook, the code values for OP36 and OP37 are incomplete and should be as follows:

For OP36,

• 4 = "does not charge"

For OP37, the code values should be as follows:

• 1 = yes, credit only
• 2 = yes, installment only
• 3 = accept both credit and installment
• 4 = no, do not accept

The following revisions are our best guesses and are not based on an examination of the hardcopy questionnaires. Users must decide whether to use these suggested corrections:

PEMITD01:

• WOMID='10 2 21 1410' and D0_COL='D6B': set D0MES2=154
• WOMID='21 418A439 2' and D0_COL='D6A': set D0MES4=3 and set D0MES5-D0MES9=missing
• WOMID='21 418A439 2' and D0_COL='D6B': set D0MES4=missing and set D0MES5-D0MES9=154
• WOMID='21 7 2A 93 2' AND D0_COL='D6B': set D0MES8=152

PEMITD02:
WOMID='21 418A439 2': set D10_2=15

PEMITD03:
WOMID='21 213A165' and D1_COL='D68A': set D1MES4=154

Date: 7/98

Revisions to OTROS Files in Community Data. The variable OTHID in the OTROS files in the Community/Provider data was found to be incorrect. It did not include the provider number and thus was the same value for all providers in the same community. The error has been corrected in the online OTROS subfiles.

There were 4 cases in the OTROS files for which PID was missing. Those providers were in communities 4 6 1 and 10 6 3; they now have PID values in the revised online OTROS files mentioned above.

Codebook Revisions. The following relate to typographical errors in the codebook. The codebook will not be revised. Please make these revisions in your own copies:

In the codebook, the descriptions for variables AM03A and AM03B, D25A and D25B in PEMITD02, and D87A and D87B in PEMITD04 are incorrect. AM03A, D25A, and D87A are time in minutes; AM03B, D25B, and D87B are time in hours. In most cases, a non-99 value is in one or the other in the pair; respondents were supposed to give time in minutes or in hours only, not in both. In some cases, the non-99 values appear to be in both variables in the pair. In these instances, it is not entirely clear whether time was reported in hours and minutes or whether the data were entered as implied by the questionnaire: the A variable is the amount of time and the B variable is the unit of time.

In the codebook, the definition for variables D15_1­D15_5 in PEMITD02 and D77_1­D77_5 in PEMITD04 is incorrect. The definition should be DID PROVIDER TAKE BLOOD, not DID PROVIDER TAKE URINE SAMPLE. A check of the questionnaire (DRU-1538/1-NICHD) shows that the question refers to taking a blood sample.

In the codebook, Appendix B, two additional codes should be added to the codes for PC18/ME14. Code 115 is "trained in nursing." Code 116 indicates people trained by the doctor at the clinic to help with laundry, cleaning, etc. Each of these codes appears only once in the data.

In the codebook, the definition for OP05 in OTROSO01 is "COURSES AND PRACTICUMS," not "curses and practicums."

Date: 11/97

The variable INFORMID was incorrect in the Community Provider file CLAVEI15 and has been corrected in the online CLAVEI15 subfiles.

Several variables were incorrectly written out in the Community Provider files CLAVEI12, CLAVEI13, CLAVEI14, and CLAVEI15. They have been corrected in the online CLAVES.

Several variables were incorrectly written out in the Community Provider files PUESTP01 and PUESTP04 and the Individual survey file PEMITE01. The errors have been corrected in the online PUESTOS and PEMITE01.

The online claveinp.sas, puestinp.sas, and pemit3inp.sas programs in the egsf/egsfraw/inptpgms subdirectory have been revised to correct errors in the input formats for a couple of variables.

In addition, a sample program showing how to read the SAS transport files, called convert.sas, has been put online in the egsf/egsfexp subdirectory.