This work is an extension of the methodology proposed by Polo (2018) to estimate the value-added of higher education institutions using mixed hierarchical linear models with error in the variables, where the parameters associated to the model were estimated by maximum likelihood and using the optim function of R. In this work an alternative methodology of estimation of the parameters of the model is developed via EM algorithm and implemented in R through functions that automate the estimation of the value-added of the institutions and the calculation of their standard errors via a bootstrap procedure. This method is applied to engineering, business administration, psychology and humanities students who took the Saber 11 test in the period 2006-2009 and the Saber Pro test in the years 2012 and 2013. Finally, when comparing both methodologies, it is evident that the estimates of the value-added using the EM algorithm are more precise than those obtained using the optim function.