Using the estat gof command in Stata, the F-adjusted mean residual test for goodness of fit for logistic regression models fitted using survey data was used to determine a final selection of variables that provided the best fit.
We then examined the spatial structure of the highest-ranked model's residuals, testing for autocorrelation and potential differences between islands or from macroalgae distribution.
The following criteria were used to evaluate the models (independent and simultaneous ones): adjusted coefficient of determination (Adj.R2), mean residual (MR), standard deviation of residuals (Sy.x), test for heteroskedasticity of residuals, test for normality of residuals, and graphical analysis of residuals.
In all models, adequacy of fit was considered using residuals, tests for homoscedasticity and Aikakes Information Criterion AICC).
Partial regression plots were inspected for outliers and heteroskedacity, and residuals tested for normality using the Shapiro Wilkes test.
The following criteria were used to evaluate the models (independent and simultaneous ones): adjusted coefficient of determination (Adj.R), mean residual (MR), standard deviation of residuals (Sy.x), test for heteroskedasticity of residuals, test for normality of residuals, and graphical analysis of residuals.
These code provisions are found to be more conservative for residual test data rather than hot-temperature test data.
The absence of residual infectivity was proven by exhaustive testing for residual infectivity in assays in Vero cells (15, 16).
Other residuals tests were performed for back-azimuth, hypocentral distance, and signal duration (depending on available information on the different datasets).
Assumptions for the regression model were tested by investigating the normalized distribution of residuals and also testing for collinearity and homoscedasticity of the data.
