Using calculated lagged values in a non-spatial regression model

[Sara H]

Greetings,

 I have a question regarding using spatially lagged values calculated in GeoDa. Working with autocorrelated crime data, I have calculated spatially-lagged values from inverse distance weights in GeoDa to account for spatial error. I wonder if I can use this dataset –that has already accounted for spatial error and has added lagged values to the original values— in a linear regression model (non-spatial error model)? In other words, would replacing new values that are the sum of original values and the lagged values, adequately reduces spatial error so that it could be treated as a non-autocorrelated dataset in a linear regression model?  

Thanks,

Sara