I found this study quite interesting, as it had been my belief that insulin responses were predominately from the intake of carbohydrates; however, the study shows some foods stimulate insulin responses more readily than others, regardless of carbohydrate content.http://www.ajcn.org/content/66/5/1264.full.pdf+html In contrast, pasta, oatmeal porridge,and All-Bran cereal produced relatively low insulin responses,despite their high carbohydrate contents. Carbohydrate wasquantitatively the major macronutrient for most foods. Thus, itis not surprising that we observed a strong correlation betweenGSs and ISs (r = 0.70, P < 0.001). However, some proteinand fat-rich foods (eggs, beef, fish, lentils, cheese, cake, anddoughnuts) induced as much insulin secretion as did some carbohydrate-rich foods (eg, beef was equal to brown rice andfish was equal to grain bread). As hypothesized, several foodswith similar GSs had disparate ISs (eg, ice cream and yogurt,brown rice and baked beans, cake and apples, and doughnutsand brown pasta). Overall, the fiber content did not predict themagnitude of the insulin response.So it would seem that insulin is not quite dependent upon the actual substrate of the food necessarily.  There is a correlation between high carbohydrate foods, but it doesn't explain why beef would have such a large insulin response.

Also, from that abstract on the first studies, it said they used a hyperglycemic clamp, which by this definition: http://en.wikipedia.org/wiki/Glucose_clamp_technique  would indicate that they either have glucose or insulin being constantly infused into this tissue already along with a meal.  Even while using the control with the clamp, how can you calculate a control on a hyperglycemic clamp on a fasting individual and compare that to foodstuffs someone else is absorbing?  I'm not buying that one for a 1.00 dollar.