How many controlled tests do you have on your product? Dr Rosen felt that 20 are needed; my experience is that perhaps ten are a reassuring guide, with at least five done independently.
How many of these have no negative controls? A negative control determines whether a product on test is effective by comparison.
Have you a list of references to support these two questions? (a) Ask to see the list, or have it sent to you. (b) How many results were negative as well as positive? Beware of all-positive results – they could have been ‘selected’; the inconclusives quietly buried.
Response frequency. How many times out of ten, does the product improve performance? 70% is satisfactory. 100% is highly suspect.
What are the co-efficients of variation? Ask them to provide those for the trials cited. These reveal the amount of variation in response over a range of conditions; all farms are different. Up to 50% should be acceptable, but the lower the better.
What dosage/method of use of the product will maximise the return on my investment with you? Ask the supplier what the return on investment he considers likely at half, and again twice, that of his recommend level.
Can you supply me with a model to predict responses to your product under my farm conditions? This is advanced exploratory work, which smaller firms because of their limitations, cannot necessarily provide, but the bigger, long-established firms should be able to do so.
From discussions with the late Dr Gordon Rosen.