That's actually a pretty big issue for surveys where I live. I live in a part of Denmark that is notorious for never using the full scale in surveys, to the point where it screws with data analysis.
Basically we'll never use 5 in a 1 - 5 score, or 10 in a 1 - 10, because what if something better came along?
I've been to training seminars and done surveys about workplace issues where we've been specifically instructed to use the full scale, because clustering on e.g. 2 - 4 will result in NOTHING, it gets removed by HR/software/analysis as average and not worth dealing with, resulting in no change.
I'd still argue that it issue is in how people use the data, but it remains a problem.
I often answer in the 2-4 range. Usually because of the way the question is phrased. If they want different answers, they should ask their questions differently.
