navigation
nav
Search
nav nav nav nav
research studies

 

 

Study Methodology

The National Study of Licensed Social Workers was designed to provide comprehensive information about the frontline professional social work labor force, the context of their practice, descriptions of services they provide, and characteristics of the client populations served.

Data were collected from 4,489 licensed social workers from forty-eight states and the District of Columbia through a mailed survey instrument. These responses were based on a stratified random sample of 10,000 licensed social workers across the U.S. Details of the sampling procedure are provided below.

Survey design. The design of the instrument was informed by extensive interviews and focus groups with practicing social workers, including a number of social workers specifically drawn from the areas of child welfare/family social work, aging, behavioral health, and health.

The core survey had four sections: Background, which included questions on demographics and education/training; Social Work Practice, which included questions on hours worked, roles, setting, practice area, and salary; Services to Clients, which included questions on tasks and caseload; and Workplace Issues, which included questions about changes in the practice of social work, satisfaction, and career plans.

Additionally, special supplements were included in the instrument for social workers who serve older adults (age 55 and older) or children and adolescents (age 21 or younger). These supplements gathered more detailed information on working with these populations.

Sampling and survey administration. A database was constructed from approximately 255,000 names of licensed social workers from state licensure and registration lists. These lists included anyone credentialed by the state as a social worker, regardless of whether the state title was licensed social worker, certified social worker, registered social worker or any other.

The list was then stratified by Census division. The U.S. Bureau of the Census recognizes nine such divisions: New England, Middle Atlantic, East North Central, West North Central, South Atlantic, East South Central, West South Central, Mountain, and Pacific. The purpose of the purpose of the stratification was to draw equal-sized samples from regions of the country that are both heavily and sparsely populated. This strategy resulted in a sample in which social workers in less-populated divisions were over-represented, which was desirable because it allowed large enough samples from each division to permit meaningful analysis of regional and rural/urban differences.

A random sample of 9,999 social workers was drawn from this master list (1,111 from each of the nine Census Divisions). The sample was then analyzed for duplicate names, which were eliminated and replaced with other randomly selected names from the same Census division.

Table 1 shows that the final sample represented approximately 4% of the master list. This represented very different proportions of the social workers in each division, however-from 8 percent of social workers in the East South Central division to 2% of social workers in the South Atlantic division.

Table 1. Sampling Rates for Census Regions for the 2004 Licensed Social Worker Survey

Census Region Total
number
Percent Number Percent of
total
New England 14,436 5.67 1,111 7.7%
Middle Atlantic 25,267 9.93 1,111 4.4%
East North Central 57,174 22.46 1,111 1.9%
West North Central 24,904 9.78 1,111 4.5%
South Atlantic 56,265 22.11 1,111 2.0%
East South Central 13,974 5.49 1,111 8.0%
West South Central 25,040 9.84 1,111 4.4%
Mountain 15,595 6.13 1,111 7.1%
Pacific 21,859 8.59 1,111 5.1%
Total 254,514 100 9,999 3.9%

Table 2 shows the response rates by Census division. The highest response rate was in the Middle Atlantic (53%) and the lowest in the South Atlantic (46%).

Table 2. Response Rates by Census Division

Census Division Total -- all mailings Total surveyed
Response rate Responses Removals
New England 476 273 1,261 48.2%
Middle Atlantic 564 115 1,183 52.8%
East North Central 471 197 1,204 46.8%
West North Central 488 113 1,067 51.2%
South Atlantic 469 190 1,205 46.2%
East South Central 501 173 1,200 48.8%
West South Central 504 62 1,135 47.0%
Mountain 521 198 1,202 51.9%
Pacific 495 210 1,191 50.5%
Total 4,489 1,531 10,648 49.2%

Survey analysis. The strategy for analysis centered on variation by demographics, degree, and sector. Only data from active social workers were used in the analyses unless otherwise specified.

A number of variables used in these analyses were created from the survey data. "Active" status was defined as working either a full-time or a part-time job in social work. "Sector", which was asked in detail, was grouped into four categories: public sector (which included federal, state, and local government and military), private non-profit, private for-profit other than private practice, and private practice. Social workers were asked to indicate all degrees they held in both social work or another field. Highest social work degree was the most advanced of the social work degrees indicated, although some respondents held a higher degree in another field than they did in social work.

Age and income were asked as categorical variables, but an estimation procedure was used to assign exact values from within each category randomly to each respondent in that category. This procedure allows some statistical procedures, such as the estimation of mean values and the use of regression analysis, which would not be possible with categorical data. This procedure also allowed the calculation of an "age at entry", which was defined as the estimated age of respondents in the year in which they reported receiving their first social work degree: the BSW (if applicable), or the MSW (if they did not hold a bachelor's degree in social work). Age at entry could not be calculated for licensed social workers who did not hold a BSW or MSW.

Data limitations. Although these data represent an important contribution to knowledge of licensed social workers, the data are not generalizable to non-licensed social workers, who may perform different functions and serve different populations. When statements are made about the percentage of social workers doing policy development, for example, the word "licensed" should always be understood even if not explicitly stated.

Return to National Study of Licensed Social Workers Readers' Guide