Overview of the Methodology
The NAWS samples and interviews workers where they are employed. Trained interviewers conduct face-to-face interviews with a national probability sample of workers employed in crop agriculture. Depending on the information needs and resources of the various Federal agencies that use NAWS data, between 1,500 and 3,600 workers are interviewed each year.
Multi-stage sampling accounts for seasonal and regional fluctuations in the level of farm employment. To capture seasonal fluctuations in the agricultural work force, the fiscal year is divided into three interviewing cycles. To capture regional variation, workers are sampled from 12 regions, which are aggregated from 17 USDA-designated regions.
There are seven levels of sampling:
- Cycle (strata);
- Region (strata);
- Single counties or groupings of counties called farm labor areas (FLA), which constitute the primary sampling unit;
- ZIP CodeTM region;
- Employer; and
- Farm worker.
Interviews are allocated proportionately to each region/cycle strata based on the amount of farm labor in that region during that cycle. In every cycle, in each region, a random sample of FLAs is selected from a roster that is constructed from the universe of FLAs using probabilities proportional to the size of a given FLA's seasonal farm labor expenditures. Similarly, counties are randomly selected in each FLA using farm labor expenditures. Within each county, ZIP CodeTM regions are randomly ordered.
In the penultimate sampling stage, a simple random sample of agricultural employers is selected from a list of employers. Once the sample of employers is drawn, interviewers contact the selected growers or contractors to obtain access to the work site.
The sampling frame of establishments (usually farms) engaged in Crop Production (NAICS 111) and Support Activities for Crop Production (NAICS 1151) is constructed primarily from the Quarterly Census of Employment and Wages (QCEW): a census of establishments that participate in the unemployment insurance (UI) program. In states where UI coverage is not universal, the sampling frame is supplemented with farm establishments drawn from commercial lists of agricultural employers, administrative data sources, internet searches, and discussions with local, knowledgeable sources.
In the final stage, workers are randomly sampled at their farm job sites. Randomly-selected workers are usually interviewed at the worksite, either before work, after work or during breaks. Respondents may also be interviewed at another location if that is more convenient.
NAWS Survey Documentation
For additional information on the NAWS methodology, consult the following documents:
- Justification for the National Agricultural Workers Survey (Shortened version of Part A of the Paperwork Reduction Act [PRA] Information Collection Request [ICR])
- Statistical Methods of the National Agricultural Workers Survey (Shortened version of Part B of the PRA ICR)
- Sampling regions
- Correspondence between NAWS and USDA Farm Labor Survey sampling regions
- Introduction to Analyzing the NAWS Public Access Data
- Sample letter to agricultural employer
- NAWS brochure