PURPOSE: To determine whether data from a professional society placement service--the Professional Bureau of the American College of Radiology--are a valid measure of the employment market. MATERIALS AND METHODS: For the United States from 1990 to 1998, the authors compared three placement service measures-the annual number of job listings, job seekers, and listings per seeker-with two presumably valid measures of the employment market-annual total jobs available (which was ascertained from surveys of hiring) and radiologist median income relative to the all-physician median. For the comparisons, both graphic displays of the data and correlation were used. RESULTS: In graphs, patterns of change were similar. The correlation of job listings, which measure demand, with total jobs, which also measure demand, was 0.84 (P =.04). The correlation of (a) job seekers, a measure of supply, and (b) listings per seeker, which involve both supply and demand, with total jobs was substantial but lower: 0.58 (P =.23) and 0.76 (P =.08), respectively. Correlation of the three placement service measures with relative income, which presumably depends on both supply and demand, was 0.80-0.88 (P <.05 for each measure). CONCLUSION: The statistical significance levels of the correlations and the pattern of findings--namely, stronger correlations among measures of the same aspect of the employment market--indicate that these placement service data are valid and reasonably accurate measures of the employment market.
PURPOSE: To determine whether data from a professional society placement service--the Professional Bureau of the American College of Radiology--are a valid measure of the employment market. MATERIALS AND METHODS: For the United States from 1990 to 1998, the authors compared three placement service measures-the annual number of job listings, job seekers, and listings per seeker-with two presumably valid measures of the employment market-annual total jobs available (which was ascertained from surveys of hiring) and radiologist median income relative to the all-physician median. For the comparisons, both graphic displays of the data and correlation were used. RESULTS: In graphs, patterns of change were similar. The correlation of job listings, which measure demand, with total jobs, which also measure demand, was 0.84 (P =.04). The correlation of (a) job seekers, a measure of supply, and (b) listings per seeker, which involve both supply and demand, with total jobs was substantial but lower: 0.58 (P =.23) and 0.76 (P =.08), respectively. Correlation of the three placement service measures with relative income, which presumably depends on both supply and demand, was 0.80-0.88 (P <.05 for each measure). CONCLUSION: The statistical significance levels of the correlations and the pattern of findings--namely, stronger correlations among measures of the same aspect of the employment market--indicate that these placement service data are valid and reasonably accurate measures of the employment market.