If you have a rating scale such as are used on product websites where users enter scores 1,2,3,4,5 (five values) and averages are (presumably) rounded to discrete intervals 1,1.5,2,2.5,3,3.5,4,4.5,5 (nine values), what is the best method for calculating lower and upper confidence limits around a measure of centrality. The LCL/UCL and mean(?) would all have to be rendered in the nine point scale 1,1.5,2,2.5,3,3.5,4,4.5,5. Thanks for your thoughts and help!