Study Description

Wisconsin Basic Needs Study, 1981-1982
Merged Data Set, Waves 1 -5


The format of the data for the Basic Needs Study is quite complex, reflecting the methodology of the time that the data were collected. The information included below is essential reading for any researcher trying to use these data. Much of this information is derived from the original project codebook introduction and other project papers. DPLS staff re-wrote this information and clarified it as was necessary.

The sample consisted of a cross-section of households throughout the State of Wisconsin, supplemented by special oversampling of households in which the State Department of Health and Social Services had a particular interest (low-income families, households with an elderly head, and female-headed households with children).

The total sample size is 1816 households.

There are 6 records per household.

Each record is 28,000 characters in length.

Each variable is 10 characters in length.

The objectives of the survey were to describe the basic needs of families, to observe the variation in needs as a function of household size and composition, region, and season, and to determine the behavioral and psychological responses to changes in the ability of families to satisfy their consumption requirements. Included were an extensive retrospective life history and behavioral and attitudinal measures.


MacDonald, Maurice and Diane Colasanto. Wisconsin Basic Needs Study [Machine-Readable data file]. 1st DPLS ed. Madison, WI : Institute for Research on Poverty [producer], 1982. Madison, WI: University of Wisconsin Data and Information Services Center [distributor], 1996. 1 data file (1816 logical records); <>; ().

Format of the Data

The merged dataset of 11,900 variables is a combination of household and person data over the five waves. The five logical data "segments" are described below:

  • Personal Interview Variables 1-5000
    • Household Variables 1-2800 (2800 variables)
    • Person Variables 2801-5000 (2200 variables)
  • Wave 2 Variables 5001-6600 (1600 variables)
  • Wave 3 Variables 6601-8200 (1600 variables)
  • Wave 4 Variables 8201-9800 (1600 variables)
  • Wave 5 Variables 9801-11900 (2100 variables)

Each of the preceding six sets of variables were written out as individual rectangularized records with a fixed record length of 28,000. Thus the data for a single household consists of six such records:

  1. Record 1 (Wave 1 Household)
  2. Record 2 (Wave 1 Person)
  3. Record 3 (Wave 2)
  4. Record 4 (Wave 3)
  5. Record 5 (Wave 4)
  6. Record 6 (Wave 5).

Records were padded out to their full length with -1 codes.

For any given wave all of the Variables were not used. A few unused Variables were re-initialized to 0 (this is an exception to the -1 rule specified below). Each Variable is 10 characters, and all appear to be right justified. The first Variable (or Word in the parlance of the Principle Investigators) in each of the five original data Segments is the Household ID.

Household ID number indicates sample:

  • 0001-3999 Area Probability Sample
  • 4000-4999 AFDC Sample
  • 5000-5999 Female-Headed Households with Dependent Children
  • 6000-6999 Aged Household Sample (Head 65 or older)
  • 7000-7999 Low-Income Household Sample

The second word is a household indicator, coded as 1 if the household is present in that wave, otherwise coded as 0 (and then all remaining variables in that segment are left as -1).

A few Variables have implied decimals (as noted in the codebooks) and they are as follows:

  1. Variable 151 in the Personal Interview - Weight - has three implied decimals
  2. Hourly Wage and Salary Rates were coded with two implied decimals "in Waves 2 - 4".

Within each of the original five wave segments there are person variables (in Wave 1 these have been separated onto the second record). Because it was known that the maximum number of people in any household across all five waves was eleven, the person data sections were set up to allow for the maximum of 11 persons. Each of these sections has a first variable that indicates whether a person of that number is present at this time, and if not, all the variables for that non-present person in that wave are set to the default of -1.

Because the order of the variables reflected the original coding procedure within each household and person grouping, the ordering is not necessarily meaningful or convenient. This has distinct implications when browsing through the codebooks, as there are logical skips of considerable magnitude among variable groupings.

Throughout the interviews some information was obtained by Person Number rather than for the household. It was this information that was inserted into the person data sections of each wave. For example, in the Telephone Waves (Waves 2 -4), Social Security was obtained not at the household level, but at the person level. However, it was also possible that the respondent would acknowledge that Social Security was received but either didn't know by whom or it was not ascertained, and that information was carried along in the household section of the data. Thus, in order to calculate total Social Security received by the household it will be necessary to sum across all persons present that wave as well as the household variable that contains information about "not ascertained" people.

For most of the continuous dollar variables, but certainly not all, an attempt was made by the Principle Investigators to make the data more easily accessible for analysis. For these variables, there is no missing data, as it is assumed to be zero. However, for each variable there is an accompanying flag that informs the researcher as to the quality of the dollar value. Thus, it is left to the discretion of the researcher as to how missing data will be handled. The definition of these accompanying flags easily identifies which variables have been handled in this manner, and the flag always immediately precedes the dollar variable.

Missing Data Information

Initially all variables were set to -1 (the Not Applicable Code). The Code of -1 is repeated throughout the data and reflects when the skip local specifies that a question is to be skipped. In the codebooks, following the text of the question, the skip logic is specified. If there is some skip logic for the question, it is contained between asterisks, and should be read as "if [statement that appears between asterisks] is true, variable will be set to -1."

In addition, a code of -2 indicates a response of "don't know" and -3 indicates a "not ascertained" response.

There are two variables that may have a legitimate negative response:

  1. Century month of birth. If a person is born before December, 1899, century month calculates to be negative.
  2. Variable 557 in the Personal Interview (Wave 1) has a legitimate response of "it depends" and this was coded as -4 to avoid confusion with any positive response.

Variable Identifiers

Corresponding to every variable in the codebooks is a wave-deck-variable identifier in the form i-jj-kk where:

  1. i denotes the wave source (1-5)
  2. jj assumes the deck numbering convention of the original data collection effort
  3. kk was the original variable order

The VARxxxx number associated with each variable in the codebook reflects the actual location of that Variable in the data record. Thus VAR0036 in the Wave 1 codebook occupies columns 361-370. All variables are right justified. There are two important exceptions to this rule:

  1. Post-project processing of the data (which included conversion from the original integer binary to ASCII) split Wave 1 into two records at Variable 2801 (see above). Researchers must re-calculate the location of those variables 2801 on as starting in Variable position 1 on record 2.
  2. When a specific group of variables is repeated (for examples see the equivalencies information that follows) only the VAR numbers of the first group are given. It is left to the researcher to deal with calculating the locations of the other groupings.

Format of the Codebooks

Although the data have been merged into one file, the codebooks are in three parts. Each codebook begins with VAR0001. Users must calculate for themselves the location of each variable based on word size (10 characters) and variable position on the record. In December of 2000, DPLS staff converted the documentation files from gif to pdf format for easy browsing.

  1. Wave 1 codebook Personal Interview (PDF, 2035K)
  2. Waves 2-4 codebook The three telephone interviews are identical thus the variables are only listed once. (PDF, 1119K)
  3. Wave 5 codebook Final Interview (PDF, 1022K)

Equivalencies Tables

Wave 1 (Personal Interview)

Record 1-Household Data

Personal Interview 
(Wave 1 but does contain some cross-wave information)

Household Demographic Data Wave 1-5  Variables 1-100

Household Data from Wave 1     Variables 101-1000

Respondent Wave 1 Job Information  
(coded for up to 16 jobs since 1976)
      Variables 1102-1119 (job 1)
      Variables 1120-1137 (job 2)
      Variables 1372-1389 (job 16)

Fertility Wave 1  Variables 
(coded for up to 16 children)
      Variables 1402-1419 (first child)
      Variables 1420-1437 (second child)
      Variables 1672-1689 (sixteenth child)

Person Demographic Variables
      Variables 1701-1800 in codebook 
	    repeat for each individual
      Maximum person number across 
	    all 5 waves of interviews is 11
      Variables 1701-1800 refer to person 
	    with person number 1
      Variables 1801-1900 refer to the person 
	    with person number 2
      Variables 2701-2800 refer to person 
	    with person number 11

Record 2-Person Data from Wave 1

Variables 2801-3000 in codebook 
  repeat for each individual
Person 1 Wave 1 Data  Variables 0001-200  
  (In Wave 1 Codebook Variables 2801-3000)
Person 2 Wave 1 Data  Variables 201-400
Person 11 Wave 1 Data Variables 2001-2200

Wave 2

Record 3

Household Data from Wave 2  Variables 1-1600 (
  Original Variables 5001-5500) 

Person Data Wave 2
Variables 501-581 in codebook repeat for each individual
Person 1 Wave 2 Data  Variable Locations 5501-5600
Person 2 Wave 2 Data  Variable Locations 5601-5700
Person 11 Wave 2 Data Variable Locations 6501-6600

Wave 3

Record 4

Household Data from Wave 3 Variables 1-437 
  (Original Variable Locations 6601-7100)  

Person Data Wave 3
Variables Numbered 501-581 in codebook repeat 
  for each individual
Person 1 Wave 3 Data  Variable Locations 501-700
Person 2 Wave 3 Data  Variable Locations 701-900
Person 11 Wave 3 Data Variable Locations 2601-2700

Wave 4

Record 5

Household Data from Wave 4 Variable Locations 1-437 
  (Original Variable Locations 8201-8700)  

Person Data Wave 4
Variables Numbered 501-581 in codebook repeat 
  for each individual
Person 1 Wave 4 Data  Variable Locations 501-700
Person 2 Wave 4 Data  Variable Locations 701-900
Person 11 Wave 4 Data Variable Locations 2601-2700

Wave 5

Record 6

Household Data from Wave 5 Variable Locations 1-952 
  (Original Variable Locations 9801-10113)  

Marital Relationship Data Variables Numbered 314 (counter) 
  and 315-324 in codebook repeated for each of up to 
  four relationships

Relationship 1 Variable Locations 315-324
Relationship 2 Variable Locations 325-334
Relationship 3 Variable Locations 335-344
Relationship 4 Variable Locations 345-354

Respondent Jobs Since 1980 
(Variable Numbers 371-385 repeated for each of up to 8 jobs)

Job 1 Variable Locations 371-385
Job 2 Variable Locations 386-400
Job 8 Variable Locations 476-490

Material Goods Matrices
Each item can have up to eight instances (unless otherwise 
  indicated). Counter variables indicates number, then 
  grouped variables follow as in this example:

Variable Location 494 Number of Cooking Stoves
Cooking Stove Variable Numbers 495-497 repeated for 
  up to each of eight items.
Cooking Stove 1 Variable Locations 495-497
Cooking Stove 2 Variable Locations 498-500
Cooking Stove 8 Variable Locations 516-518


Counter Location 519
Instance 1 Locations  520-522
Instance 8 Locations  541-543


Counter Location  544
Instance 1 Locations  545-547
Instance 8 Locations  566-568


Counter Location 569
Instance 1 Locations 570-572
Instance 8 Locations 591-593

Clothes Washers

Counter Location  594
Instance 1 Locations 595-597
Instance 8 Locations  616-618

Clothes Dryers

Counter Location 619
Instance 1 Locations 620-622
Instance 8 Locations 641-643

Sewing Machines

Counter Location  644
Instance 1 Locations 645-647
Instance 8 Locations 666-668

Central Air Conditioners

Counter Location 669
Instance 1 Locations 670-672
Instance 8 Locations 691-693

Separate Air Conditioners

Counter Location 694
Instance 1 Locations 695-697
Instance 8 Locations 716-718


Counter Location 719
Instance 1 Locations 720-722
Instance 8 Locations 741-743


Counter Location 744
Instance 1 Locations  745-747
Instance 8 Locations  766-768


Counter Location  769
Instance 1 Locations 770-772
Instance 8 Locations 791-793


Counter Location  794
Instance 1 Locations  795-797
Instance 8 Locations  816-818

Lawn Mowers

Counter Location 819
Instance 1 Locations 820-822
Instance 8 Locations 841-843

Garden Tractor

Counter Location 844
Instance 1 Locations 845-847
Instance 8 Locations 864-868


Counter Location  869
Instance 1 Locations  870-872
Instance 8 Locations  891-893


Counter Location  894
Instance 1 Locations 895-896
Instance 6 Locations 905-906


Counter Location  907
Instance 1 Locations  908-909
Instance 6 Locations  918-919

Motorized Campers

Counter Location 920
Instance 1 Locations 921-922
Instance 6 Locations 931-932


Counter Location  933
Instance 1 Locations 934-935
Instance 6 Locations 944-945

Cars, trucks, vans

Counter Location  946
Instance 1 Locations 947-952
Instance 9 Locations 995-1000

Person Data Wave 5
Variables Numbered 1001-1051 in codebook for Wave 5 
  repeat for each individual
Person 1 Wave 5 Data  Variable Locations 1001-1100
Person 2 Wave 5 Data  Variable Locations 1101-1200
Person 11 Wave 5 Data Variable Locations 2001-2101