P-value is too high - Githubissues

The crime rate for these 20 units does show spatial autocorrelation:

> lm.morantest(lm_null, listw=lw, alternative="two.sided")

    Global Moran I for regression residuals

data:  
model: lm(formula = ratecrimes ~ 1, data = wpom3)
weights: lw

Moran I statistic standard deviate = 2.5909, p-value = 0.009571
alternative hypothesis: two.sided
sample estimates:
Observed Moran I      Expectation         Variance 
      0.34570598      -0.05263158       0.02363659

However, when the linear model has explanatory variables, the residual spatial autocorrelation is removed, probably because one or more of the independent variables has a similar map pattern to the crime rate:

> lm.morantest(lm_obj_pre, listw=lw, alternative="two.sided")

    Global Moran I for regression residuals

data:  
model: lm(formula = form_pre, data = wpom3)
weights: lw

Moran I statistic standard deviate = -0.22291, p-value = 0.8236
alternative hypothesis: two.sided
sample estimates:
Observed Moran I      Expectation         Variance 
     -0.14407690      -0.11006321       0.02328247

Look at:

cor(model.matrix(~ -1 + ratecrimes + log(density) + avgsal + unemp + ssusers, wpom3))

The variables are highly correlated; are they all needed?

wikpur / Spacial-Econometrics-Project

P-value is too high #22