Linear regression seeks to control the influence of observed characteristics on the outcomes.
It assumes a linear relationship between outcome (e.g. the investment in R&D) and other characteristics of the firm (e.g. sector, age, size etc., including the granting of aid). It is possible to see linear regressions as a linear approximation of more complicated relationships. Linear regressions can be seen as general-purpose techniques and are used in many different evaluation contexts.