I really need help with SPSS (v22). Im analyzing the banking system of my country. In my sample dependent binary variable [market_default] has two states «0» - bank continues operations and «1» - bank in default. As you can easily guess, the measurement type "0" significantly outweigh measurement type "1" (number of live banks more than bankrupt).
I believe it's possible to do the following - to construct a set of logit regressions based on all observations of type "1" (default) and the observations of the type "0" (normal), with some observations such as "0" will be randomly extracted from the sample. So I can raise the average defaulted industry from 1% to 5-7% and get more robust estimates.
Question - how can I get the program to include all f the estimates type «1» in regression and random (bootstrapped) estimates type «0».
Thank you very much for your answers!