It is also possible that they were looking for 'questions that indicate cluster membership'. RWA could be defined as "tendency to trust in authority, value obe

It is also possible that they were looking for 'questions that indicate cluster membership'. RWA could be defined as "tendency to trust in authority, value obeying rules, and dislike people who break rules". It's a bad name, as someone in a left wing country with high RWA would tend to disapprove of right wing protesters for instance. So in designing the test you put together a huge list of questions that vagually relate to RWA, collect answers, and then look for ones that tend to cluster together. 'People who answer yes to this tend to answer yes to these others'. You then remove the ones that show poor correlation, and you have a test.