Hi Eleanor

May I suggest that there is a barrier to evaluation based on the value that everybody must be treated the same? I would suggest that your article could elaborate the need for experimental evaluation of innovations using a statistical design. This would require that there:

• are experimental and control groups;
• the innovation is so well defined that others can introduce it;
• that the innovation is done the same way each time;
• and, that the participants go through the procedures that are expected of them.

Just a thought.

Stu

Updated 12/09/03