COMPUTER OPERATIONS REQUEST #337a Date: January 1983 Purpose: Create a variable Size of Residence in Hundreds (SPL100). This variable will be added to the WLS Master Tape cols. 448-552 6/83. Documentation of SPL100 1. SPL100 General Description SPL100, SIZE of 1975 Place of Residence in 100's (l970), records l970 Census population counts, in hundreds, for the l975 places of residence, in the State of Wisconsin, for members of the l975 survey. Special codes are assigned for (1) respondents whose l975 place of residence had a population of less than 1000 in l970, (2) respondents who were residents of Wisconsin but had not been assigned a place code, (3) respondents who were not residents of Wisconsin in l975, and (4) respondents for whom there was no known l975 address. 2. Extracting Population Counts and Geographic Codes from the 1970 Census Files The first step in creating SPL100 was to extract population counts and geographic codes from the l970 Census files for places in Wis- consin. The geographic codes used in Census files are described in the U.S. Bureau of the Census, 1970 Census of Population and Housing, Final Report PHC(R)-3; Geographic Identification Code Scheme, North Central Region (April, 1972). Population counts for Places in the U.S. are released by the Bureau of the Census as Second Count Summary Tapes, File B: Second Count Population and Housing Summaries for Selected Census areas, with one file for each state. For the pur- pose of creating SPL100, the Second Count Summary File B for Wisconsin was available. A complete description of this file can be found in the National Data Use and Access Laboratories (DUALabs) Technical Doc- ument No. ST-2: 1970 Census of Population and Housing Second Count Summary Tape (August, 1972). The DUALabs program MOD-4 (Instruction Booklet for the MOD-4 Program, Revision 6 version, January, l973) was used to create an extract of the 1970 Wisconsin File B containing geographic codes and population counts for Places in Wisconsin. The geographic codes extracted were (1) 1970 State code ( = 55 for Wisconsin), (2) 1970 County code, and (3) 1970 Place code. Other components of the Census Geographic Identification Code Scheme code were also extracted but were not subsequently required for the creation of SPL100. The structure of the Census file is such that for each Place total population counts are broken down by race and sex. Getting the total population for each Place thus required summing counts across race and sex categories with the MOD-4 "MATH" procedure. In addition, the total count was divided by 10 to create a "Size of Place in 10's" code - MOD-4 does not have the facility to correctly round the product of its division function, so a count reflecting Size of Place in 10's was output. This count was transformed into a cor- rectly-rounded Size of Place in 100"s count in a later step in creating SPL100. In addition, two of the detailed race by sex population count categories were extracted in order to check for the Census Bureau's data-supression codes - none were found. Census Extract: INFILE (Second Count File B: Wisconsin): POP702B.WIS Tapes: DE7751, DE 7752 OUTFILE (on disk): [236006]SPLCNT.DAT Record Length: 72 Blocking Factor: Unblocked Records: 309 Fields DUALabs Data Descriptors Label 1-8 G01001 l970 State Code 9-16 G05001 1970 Place Code 17-24 G14001 Place Description Code 25-32 G15001 Specified City with Rural Territory Code 33-40 G20001 Place Size Code 41-48 001001 Count of White Males 49-56 001002 Count of Negro Males 57-64 TOTPOP Computed count of Total Population for 1970 Place: Sum of Data Descriptors 001001 thru 001018 65-72 SPL10 Size of 1970 Place in 10's: TOTPOP/10 with decimal remainder dropped 3. Extract of WLS file KSIB.DAT, an extract of the WLS file PSIB.DAT, was created using the CDE EXTRACT program. The PSIB.DAT fields extracted, along with their location in KSIB.DAT are as follows: PSIB.DAT Fields Variable KSIB.DAT Fields 1-6 IDSWL 1-6 107-115 GEOG57 7-15 256-267 GEOG64 16-27 429-437 GEOG70 28-36 439-447 GEOG75 37-45 454-456 MCD70 46-48 457-459 MCD75 49-51 INFILE (WLS): PSIB.DAT Tape: DE 7671 OUTFILE: KSIB.DAT Tape: DE 8048 Record Length: 51 Blocking Factor:100 Records: 10,317 4. Matching Place Size Counts to WLS Respondents Each respondent's 1975 place of residence was, according to Appendix E: Geographic Location Codes in the WLS documen- tation, coded by state, county, and place using the Census Geo- graphic Identification Code Scheme. These three geographic divisions are recorded as GEOG75 in columns 439-447 of PSIB.DAT. Although Appendix E does not indicate the arrangement of the three codes within GEOG75, an examination of 2 blocks of data in the extract KSIB.DAT showed that GEOG75 contains, in order, the 2 digit STATE code, the 3 digit COUNTY code, and the 4 digit PLACE code. Because the geographic codes in the Second Count Summary B file are also GICS codes, and because Place codes in the GICS are unique within state, a direct match between the WLS GEOG75 STATE and PLACE codes and the Census extract STATE and PLACE codes was made. The value of SPL100 assigned to each respondent is determined as follows: A. Respondents for whom there was no known 1975 address were, according to WLS documentation Appendix E, coded with 0's in the State, County, and Place code fields in GEOG75. The matching procedure checked each respondent's 1975 State code field for the Wisconsin code "55". Where the State code did not equal 55, the respondent was assigned the value "99999" for SPL100. B. Respondents who were not residents in Wisconsin in 1975, but for whom a valid (i.e., non-zero) State code existed, were also coded "99999" for SPL100. That is, all cases where the 1975 State code was not equal to 55 were coded with 9's in SPL100. To differentiate respondents of this type from those of type A above, thus requires using their 1975 State code in GEOG75. C. Respondents who were residents of Wisconsin in 1975 (State code = 55), but who did not have a valid code for 1975 Place (Place code = 0) were assigned the code "99998" on SPL100. According to Appendix E in the WLS documentation, respondents with State code = 55 and Place code = 0 were those who were residents of Wisconsin but for whom a GICS PLACE code could not be found - their place of residence is not defined as a PLACE in the GICS. D. Respondents who were residents of Wisconsin in 1975 (State code = 55) and who had a valid (i.e., non-zero) 1975 Place code, but for whom no matching Place code was found in the Census extract SPLCNT.DAT were assigned the value "9" on SPL100. These cases occur because the Census Second Count Summary B files do not report population counts for Places of less than 1000 popu- lation. In the matching procedure it was assumed that all res- pondents with a valid Wisconsin Place code who were not matched to a Wisconsin Place code in the Census extract file were residents of Places with less than 1000 population. A subsequent check against the published Census population counts for Places in Wisconsin showed that in every instance the Places for which SPL100 was coded 9 were in fact Places with less than 1000 pop- lation. E. Respondents who were residents of Wisconsin in 1975 and whose 1975 Place code was matched by a Place code in the Census extract SPLCNT.DAT, were assigned a value in SPL100 reflecting the 1970 population count for the Place divided by 100 (In the actual procedure the value of SPL100 was determined by dividing the variable SPL10, Size of Place in 10's, by 10 and rounding). 5. Output from Matching Procedure SPLSIB.DAT was created as the output file from the matching procedure. Each record (case) in KSIB.DAT was transfered intact to SPLSIB.DAT with the new variable SPL100 added to the immediate end of each record. Following SPL100 was also added a counter variable F1, coded as follows: If SPL100 = 99999, F1 = 1: Missing or Non-Wisconsin 1975 State Code If SPL100 = 99998, F1 = 2: Missing Place code for a Wisconsin Resident If SPL100 = 9, F1 = 3: Size of Place less than 1000 for a Wisconsin Resident If SPL100 = 10-7171, F1 = 0: SPL100 contains 1970 Population count, in hundreds, for 1975 Place of Residence 6. Format of SPLSIB.DAT Tape: DE 8049 Record Length: 57 Blocking Fact: 100 Records: 10,317 Variable Position IDSWL 1-6 GEOG57 7-15 GEOG64 16-27 GEOG70 28-36 GEOG75 37-45 MCD70 46-48 MCD75 49-51 SPL100 52-56 F1 57 7. SPL100 Codes Code Label 9 Size of 1975 Place of Residence is less than 1000 in 1970 10-7171 Actual Size of 1975 Place of Residence in 1970, in hundreds 99998 1975 Place code = 0: Respondent's 1975 place of residence is not a Census-defined Place (under GICS coding system) 99999 1975 STATE code not = 55: Respondent was not a resident of Wisconsin in 1975, or did not have a known address in 1975 8. Tests of SPL100 A. For all cases with SPL100 = 99999: not residing in Wisconsin in 1975 or not ascertained (1975 STATE CODE not = 55) Frequencies of variables STATE75 and PLACE75 checked. None of the cases have STATE75 = 55. B. For all cases with SPL100 = 99998: PLACE75 Code = 0 Frequencies of STATE75 and PLACE75 checked. All cases are coded STATE75 = 55 and PLACE = 0. C. For all cases with SPL100 = 9: PLACE75 less than 1000 in 1970 Frequencies of PLACE75 checked. These codes were all checked against 1970 Census table for population of places in Wisconsin. All PLACE75 codes with SPL100 = 9 had less than 1000 total population in 1970 D. A selection of "valid" PLACE75 codes (27 in all) were selected and crosstabulated with SPL100, then checked against 1970 Census population totals for Places in Wisconsin. All the PLACE75 codes examined have the correct SPL100 code. SPL100 will be added to the WLS master tape 6/83. On the input tape SPL100 (cols. 52-56) uses codes 9-7171 as valid data codes. On the new WLS Master SPL100 will appear in cols. 448 - 452. The missing data codes will be: 77777 = Wisconsin Resident with no Place Code 88888 = Inappropriate, Not a Wisconsin Resident 99999 = Not Ascertained (GEOG75 = 999999999) COMPUTER OPERATIONS REQUEST #337 ADDENDUM I August, 1983 Updating Size of Place Variables See COR #344 for an update of SPL100 (columns 448-452) to include non-Wisconsin residents and the addition of RSPL100 in columns 460-464 (source code = GEOG70) and XSPL100 inn columns 2559-2563 (source code = GEOG77)