Some of our variables were highly correlated, which can lead to inflated standard errors and unstable regression coefficients. We use the Gram-Schmidt procedure to orthogonalize highly correlated variables, which partials out the common variance and creates transformed variables that are uncorrelated with each other.(Hiatt, S. R., Sine, W. D., & Tolbert, P. S. , 2009; Hiatt, S., & Park, S. ,2013; Galunic, C., Ertug, G., & Gargiulo, M. ,2012; Pollock, T. G., & Rindova, V. P. ,2003; Sine, W. D., David, R. J., & Mitsuhashi, H. ,2007;)

Note that one technique that does not work in offsetting the effects of multicollinearity is orthogonalizing the explanatory variables (linearly transforming them so that the transformed variables are uncorrelated with each other): By the Frisch–Waugh–Lovell theorem, using projection matrices to make the explanatory variables orthogonal to each other will lead to the same results as running the regression with all non-orthogonal explanators included.