U. S. Science Parks: The Diffusion of an Innovation and Its Effects on the Academic Missions of Universities

                                             

 

 

28

 In 27 of 29 parks we could identify the year the park was formed using information from the Internet 

and from AURRP (1997). 

---

  

20 

not perfectly mirror the population sample of 88 universities in terms of the selected key 

characteristics.  To control for differences in the probability of responding to the survey, we 

estimated the probability of responding and completing the survey, that is, the probability of 

selection into the sample of 29, prob8829.

29

  That probability is then used as a control variable in 

equation (12).

30

  We believe that this variable is doing more than simply controlling for the effect 

of a correlation in random errors in the model of response and complete models of the provosts’ 

perceptions about the effects of science parks on academic missions.  Our model of perceptions 

is exploratory and unlikely to be complete with just the variables other than prob8829.  We view 

the variable prob8829 as capturing substantive effects of the complete model that otherwise 

would be left in the error term and that are related to the probability of responding to the survey. 

Table 6 shows the econometric results for the six ordered probit models to assess the 

determinants of inter-university differences in the impact of science park relationships on the 

academic missions of universities.  The specifications presented are for the parsimonious models 

that include (apart from the response control) only the explanatory variables that had coefficients