Past research on labour-market skills shortages indicates that employers report skills shortages or hard-to-fill vacancies for a variety of different reasons. Nevertheless, there is some consensus that skills-shortages analysis needs to examine such shortages within the context of the local labour market in order to understand the labour-market dynamics and structural factors that affect the propensity for unemployed people to fill 'skills shortage' vacancies. The traditional approach has been to utilise qualitative analysis and case studies. In contrast, in this paper we undertake a multivariate probit analysis of employer perceptions of skills shortages utilising a subregional dataset from a survey of Dorset employers undertaken in 1998. On a general level, we demonstrate the complexity involved in attempting to measure skills problems using the responses of employers to standard surveys. The key findings of the probit analysis are that: firm size is a significant determinant in skills-deficiency perception, growing firms have a higher skills-shortage perception, and reported perceptions of skills deficiencies vary significantly according to the position of the respondent in the organisation. This reinforces the message that great care needs to be taken when analysing measures of skills deficiencies that are derived solely from employer surveys at national or subregional level.