We discuss measurement and propagation of errors in principal component regression. The statistical literature on measurement error models is surveyed, and used as a framework for modelling propagation of error in a regression context. The starting point for the analysis will be on quantifying the error committed while fitting the regression model. Numerical studies
will be conducted so to account for the magnitude of the error, and the distribution of the implied error will be accounted for analytically for cases where this is feasible. The project is motivated by an ongoing project with our industrial partner Unilever.