After reading
through the responses LM developed a coding frame.
Responses were then coded by LM and AF with an
inter-rater reliability of 0.86 (kappa). Any discrepancies
were discussed and resolved. Where multiple reasons
were given multiple codes were allocated