I am wondering if anyone here has ever dealt with spatial autocorrelation using Logistic Regression in GIS.

In the literature I have read so far, sometimes the issue is not even addressed. In other instances, the authors used the geographic coordinates as covariates. For example, quoting from Hu, Z., & Lo, C. P. (2007). Modeling urban growth in Atlanta using logistic regression. Computers, Environment and Urban Systems, 31, 667–688: "The second step was including spatial coordinates of data points into the list of independent variables. Spatial autocorrelation can be alleviated to some extent by attempting to introduce location into the link function to remove any such effects present (Bailey & Gatrell, 1995). For example, spatial coordinates of observations might be introduced as additional covariates, or to classify regions in terms of their broad location and treat this classification as an extra categorical explanatory factor in the model."

At the best of my understanding, the latter approach is termed "autocovariate" modeling by: F. Dormann, C., M. McPherson, J., B. Araújo, M., Bivand, R., Bolliger, J., Carl, G., … Wilson, R. (2007). Methods to account for spatial autocorrelation in the analysis of species distributional data: A review. Ecography, 30(5), 609–628. http://doi.org/10.1111/j.2007.0906-7590.05171.x

I would like to know your opinion on the issue, and what approach you happened to use.

Similar questions and discussions