Understanding Society User Support

I am carrying out an event study project using an unbalanced pooled panel dataset made up of Waves 1-14 of the UKHLS. I have created a binary variable which equals 1 for individuals who are in their last wave, and 0 otherwise. I then tested whether those that leave the survey after the current wave they are in have different characteristics to those than remain, and found that those who are about to leave have significantly different job characteristics- one of my key outcome variables in the event study. I then interacted the attrition binary variable with wave, with this telling me that attrition varies across waves unequally. My aim is to control for this discrepancy in the outcome variable bias. I have read the documents about using weights provided by Understanding Society, but I am still unsure which weighting variable and method would work best for my project. I have attempted some methods, including using indpxus_lw, but this varies within PIDP, so it doesn't work. Any help would be appreciated.