second most frequent

Hello! I have a data set of 1000 individuals (rows) and (for this example) 18 likert-scale 5-item item responses in the columns. So the data for one person might look like this: SubjectA1 1 1 2 1 2 3 5 3 4 1 1 1... etc. I want to generate 6 new variables. 3 for each of the most common 3...