Global Intercept Test

The intercepts also can be checked from a global perspective. First, write out the full model 

Looking at the model breakdown in the global coincidence test, we see that the variables that serve the intercept, other than bo, are ba, bs, b4, and bs. In order for the intercepts to meet at the same point, then b2, bs, b, and bs all must be equal to 0. Let us determine if the intercepts are all 0, using the six-step procedure. 

Because Fc = 69.329 Ft = 2.09, reject //q at a = 0.05. The equations (two products atz2, Z3, Z4,—anatomical test sites—equals six equations) do not have the same intercept. Remember, the abdominal test site is evaluated when Z2 = Z3 = Z4 = 0, so it is still in the model. To find where the intercepts differ, break the problem into three two regression equations at each of the three anatomical areas. Because the slopes were not parallel, nor the intercepts the same, the full model (Table 9.18) is the model of choice, at the moment. 

Determining the confidence intervals in indicator, or dummy variable analysis, is performed the same way as before. 

Sbi is the standard error of values around ft, and k is the number of ft values, not including bo-For example, looking at Table 9.18, the full model is 

---