The Workforce Investment Act of 1998 shifts decision-making authority for funding of local job training programs from the federal government to state and local boards, which will need local decision-making tools to inform policy. One such tool is a method proposed by S. Goetz and D. Debertin to estimate demand for educational attainment at the county level utilizing nonlocal secondary data on employment patterns. The method focuses analysis on the demand of private businesses and makes possible a simple forecast of future demand. This paper estimates local demand for educational attainment in two West Virginia counties using the Goetz-Debertin (nonsurvey) method and compares results with those from a survey of local businesses. Two assumptions of the nonsurvey method are fixed proportions of occupations per sector and fixed proportions of educational attainment per occupation. Results indicate that differences in the distribution of businesses by size and by subsector across counties might affect final demand, but the nonsurvey method does not take this into account. Indications that proportions of input might change over time suggest that the nonsurvey method should use annual data sets such as the Current Population Survey. Conflicting with the nonsurvey method's second assumption is the finding that certain occupations in different sectors differ in demand for educational attainment. The study also found that the West Virginia counties' proportion of top-level occupations was substantially below national averages, suggesting underinvestment in education, and that businesses saw a real need for more computer training. (Contains 15 references, many data tables, and the business survey questionnaire.) (SV)