Fitting spatial autoregressive logit and probit models using Stata: The spatbinary command

Abstract

Starting from version 15, Stata allows users to manage data and fit regressions accounting for spatial relationships through the sp commands. Spatial regressions can be estimated using the spregress, spxtregress, and spivregress commands. These commands allow users to fit spatial autoregressive models in cross-sectional and panel data. However, they are designed to estimate regressions with continuous dependent variables. Although binary spatial regressions are important in applied econometrics, they cannot be estimated in Stata. Therefore, I introduce spatbinary, a Stata command that allows users to fit spatial logit and probit models.

Death by the minute: Inequities in trauma care for victims of firearm violence

“…We then performed a spatial autoregressive logit model and assessed the residuals for spatial autocorrelation. 15 Finding no residual spatial autocorrelation (index < −0.001, p = 0.753), this became the final model. Statistical significance of the final model was defined as p ≤ 0.05 for multivariable assessment.…”

“…To address this, a spatial weights matrix for the shooting locations was created using inverse distance weighting and spectral normalization. We then performed a spatial autoregressive logit model and assessed the residuals for spatial autocorrelation 15 . Finding no residual spatial autocorrelation (index < −0.001, p = 0.753), this became the final model.…”

The distribution and conservation of areas with microendemic species in a biodiversity hotspot: a multi-taxa approach

“…We used spatial autoregressive logit regression models to test the hypotheses of associations between the five local attributes simultaneously and the AM probability of each taxonomic group. Models were built using the non-linear two-stage least squares (N2SLS) estimator in the spatbinary command in Stata ( Spinelli, 2022 ). Spatial models differ from aspatial models because they consider and model the spatial relationships and dependencies among data points by considering their geographic distance from each other.…”

“…An inverse geographic distance matrix generated using the spmatrix command in Stata ( StataCorp, 2017 ) was used in the model, and the models’ rho coefficients were used to assess spatial autocorrelation in the dataset. Further, Hansen’s test for overidentification was used to evaluate whether the number of explanatory variables was greater than the number of parameters to be estimated ( Spinelli, 2022 ). The coefficients of a spatial logit regression show the direction (positive or negative) of the relationship between each attribute and AM but not the attribute’s impacts on AM probability.…”

“…The coefficients of a spatial logit regression show the direction (positive or negative) of the relationship between each attribute and AM but not the attribute’s impacts on AM probability. To assess such impacts, we used the command spatbinary_impact in Stata ( Spinelli, 2022 ) to generate the elasticities ( i.e ., the percent variation in the response variable in relation to the 1% variation in an explanatory variable if the rate remained constant) of each explanatory variable of each model. Impacts can be direct, indirect, and total, where the former measures the hexagon’s predicted contributions to its probability of a positive outcome ( i.e ., being an AM), whereas the indirect impact measures the predicted impact of the other hexagons’ contributions to a hexagon’s probability of being an AM and the total impact is the sum of direct and indirect impacts.…”