Dear members of Statalist,
I am writing here regarding a dilemma that I have. I have a dataset from a survey conducted for the years 2014 2016 2018 and I would like to use especially a question, which is replicated in 2014 and 2018, but unfortunately not in 2016. I have managed to find a question in the 2016 survey that is similar to the meaning of the other one, but I am not sure whether I can modify it in order to use it in my regression. Specifically, the question of 2014 and 2018 asks whether the individuals think that a certain group is good to the country and the responses are : I totally agree, I agree, neutral, I disagree and I completely disagree, while the one from 2016 asks the individuals whether they think that the certain group is an advantage to the country and the respective options are: clearly an advantage, an advantage, neither, disadvantage, clearly disadvantage. As you can see, the content is very similar but yet they differ. If I could, I would just let it out but it is very important for my paper to capture the effects of this variable. When I asked a professor whether I can modify these questions, his suggestion was that I can, but I should express a very strong motivation in my paper. The research I have done so far is rather contradictory, some say I can do it and some not. I feel quite helpless and this forum is my last resort. Could you please give me some suggestions? Or maybe some literature that could help me overcome this difficulty?

I am grateful for your time and for your response.

Best Regards,
Lola Brahimi