Thanks for the various input. With further delving a number of additional issues arose with this measurement. In the end, I recommended the use of the higher calibration (finish constant). The courses (a marathon, half and 10km were measured in the one day, hence the long gap between pre and post calibrations) have been extended by the appropriate amounts.
I did discover that an inexperienced local measurer also joined in the exercise. This person had been trained by me at an IAAF RDC seminar in Jakarta in May. At first I thought his data would be a great help, particularly as his pre and post calibrations were almost identical. However he measured the courses to be significantly longer than the official measurer. After further examination and more questioning, I discovered I couldn't rely on his data as it would seem that he was at times distracted and didn't always follow the shortest possible route.
I approved these courses for just the one year and will insist that they be re-measured prior to the 2005 event.