I have a data set with a variable where income tax bands are coded as values. For example, 1 corresponds to the lowest income tax bracket. This means the codes are discrete integer values. I'm using this data to find the median tax bracket based on different filtered cases. In one case, the median is 5.5. What is the appropriate way to deal with this if all I want is a class interval for the median? I don't want an estimation for the exact value, so I don't need to use the formula for estimating the median of grouped data.
The other information I have is that the mean for this variable is > 5.5, but I don't know if that is a robust reason for rounding up to 6 for the tax band. I also have midpoints and upper and lower bounds for all the tax bands, including 5 and 6.