Evaluating refactorings for spreadsheet models

• We propose the first refactorings for spreadsheet models, five in total.
• We implement these refactorings in a bidirectional setting.
• We present an empirical study on the impact of end-user productivity.
• We study the effectiveness, efficiency, and user satisfaction of these refactorings.
• We discuss the results of our study.

Software refactoring is a well-known technique that provides transformations on software artifacts with the aim of improving their overall quality.We have previously proposed a catalog of refactorings for spreadsheet models expressed in the ClassSheets modeling language, which allows us to specify the business logic of a spreadsheet in an object-oriented fashion.Reasoning about spreadsheets at the model level enhances a model-driven spreadsheet environment where a ClassSheet model and its conforming instance (spreadsheet data) automatically co-evolves after applying a refactoring at the model level. Research motivation was to improve the model and its conforming instance: the spreadsheet data.In this paper we define such refactorings using previously proposed evolution steps for models and instances.We also present an empirical study we designed and conducted in order to confirm our original intuition that these refactorings have a positive impact on end-user productivity, both in terms of effectiveness and efficiency.The results are not only presented in terms of productivity changes between refactored and non-refactored scenarios, but also the overall user satisfaction, relevance, and experience.In almost all cases the refactorings improved end-users productivity. Moreover, in most cases users were more engaged with the refactored version of the spreadsheets they worked with.