Collating and interpreting employee comments

We run a number of employee surveys which involves getting anonymous feedback from employees on a range of topics affecting their experience of the workplace. The feedback is both qualitative (open-ended comments) and quantitative (numerical scores). The comments are very useful in that they provide some context to the numerical scores, so we like to pull out all the comments relating to particular items being measured.
We also like to analyze the comments by identifying the underlying theme for each comment and then grouping all comments together that have the same theme.

The job here is to read every comment, identify the theme for that comment, and then place all comments for a particular theme together. Many of the comments require some interpretation, so it would be an advantage to have an understanding of the corporate world (i.e. having worked for a large company before). The deliverable is then a document containing a number of themes, with the relevant comments underneath each theme.

The time-frames are often quite tight, so quick turnaround times are an advantage.
The cost per hour is also a big factor in making this hiring decision.