- Oversimplified models of biochemistry (way, way more complex than insulin = fat) - Sample size of 2 trials (leaving one result) - No measure of insulin (not a one-to-one relation to blood glucose, also released in complex cycles) - No control of meal time (blood glucose varies) - No control of meal contents (assuming both meals are equal in content) - No control of fasting time (blood glucose varies)
I have had to treat diabetes in cats so I got quite proficient with measuring blood glucose levels. Even when you control every single factor as much as possible (animal eats the same amount, at the same time, every day), blood glucose varies wildly. Non-diabetics have the same massive variance, as I have done blood glucose curves to myself and to others.
This experiment is unfortunately pretty much useless. The body is a very complex organic machine with lots of parts that evolved independently of each other. There aren't neat little one-to-one relationships that can be tested easily. It is full of genes that modify genes that modify genes that secrete hormones that modify pathways and brain patterns and so on.