Testing linearity in the logit using the Box-Tidwell transformation in SPSS (Part 1 of 2)

This video provides a general overview of how to use the Box-Tidwell transformation when testing the linearity in the logit assumption when performing logistic regression. This procedure is described in numerous textbooks with several references provided below: Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th ed). Los Angeles: Sage. Osborne, J.W. (2015). Best practices in logistic regression. Los Angeles: Sage. Tabachnick, B.G., & Fidell, L.S. (2013). Using multivariate statistics (6th ed.). New York: Pearson. IMPORTANT CAVEAT NOT ADDRESSED IN THE VIDEO: Make sure the variables you are including in your Box-Tidwell transformation have no negative and/or zero values (such as what would be the case with a mean-centered predictor). If the variable does have negative and/or zero values, then the natural log of those particular values will be undefined. The result when performing the Box-Tidwell transformation will be a transformed variable in the dataset that contains values that are missing. In cases where your original predictor have negative and/or zero values, you should compute a linear transformation of that variable so that no value on it is equal to or less than zero (see steps below). This is the variable you should apply the Box-Tidwell procedure to. Part 2 video demonstrating the issue and strategy for addressing it:    • Testing linearity in the logit using the B...   One suggestion for performing the linear transformation of the original variable is to: 1) Find the minimum value on that original variable in the dataset. 2) Use the compute function to compute the linear transformation in the following way: Add 1+the absolute value of the minimum value on the variable to all values on the original variable. This should produce the newly transformed variable with a minimum value of 1. 3) Apply the Box-Tidwell procedure to the variable computed in Step 2 to yield a newly transformed variable for inclusion in the logistic regression to test the assumption of linearity of the logit. ---If you have multiple predictors with zero and/or negative values you will follow steps 1-3 for each predictor. You can then test for linearity in the logit by adding the final transformed variables into the logistic regression alongside the transformed variables computed in Step 3. The SPSS data used in this video can be downloaded here: https://drive.google.com/file/d/1K7Tu...