Jane Horgan's main area of research is survey sampling and estimation, with particular emphasis on developing strategies for skewed populations with rare incidence of the study variable. These populations are referred to as "non-standard mixtures of distributions" and occur in many disciplines, notably accounting, finance, medicine, and technology production. The difficulty inherent in analysing data from such distributions is that estimators based on normal distribution theory are unreliable, and improved selection and estimation strategies are necessary. Of particular interest is the analysis of accounts in financial auditing where skewed distributions are the norm, and where existing strategies such as the Stringer, cell and moment bound estimates are too conservative to be of practical use.

In previous work, Prof. Horgan has developed new selection techniques for finite populations with unequal unit sizes. She is currently working on the use of Edgeworth expansion theory to reduce the conservativeness of the Stringer bound, on developing Bayesian strategies for application to rare-incidence populations, and on devising stratification methods for application to skewed populations with particular emphasis on sales tax audits. Work to date has been supported by Enterprise Ireland's International Collaboration funding, and by IRSCET. Future work in this area might involve developing procedures using resampling and empirical likelihood methods for application to non-standard mixtures of distributions.