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

Science Parks in the United States from 1951-1998

0

2

4

6

8

10

12

14

16

1951

1954

1957

1960

1963

1966

1969

1972

1975

1978

1981

1984

1987

1990

1993

1996

Year

Number of Science Parks

 

 

 

 

---

  

35 

 

Table 1.  Gompertz Survival Time Model of the Diffusion of Science Parks

a 

_________________________________________________________________________________________ 

Explanatory Variable 

Hazard Ratio Robust 

Std. 

Error 

 

z

b

 

Prob > |z|

b

 

______________________________________________________________________________________ 

 

Medical Center 

   1.93 

    0.519 

      2.45 

 0.014 

 

      t1 

 

   1.74 

    0.467 

      2.08 

 0.038 

 

      t4 

 

   0.649 

    0.157 

     -1.79 

 0.073 

 

     South 

 

   1.36 

    0.302 

      1.37 

 0.170 

 

   Northeast   

   1.61 

    0.465 

      1.66 

 0.097 

______________________________________________________________________________________ 

 

     gamma 

 

   0.180 

    0.0215 

      8.35 

 0.000 

______________________________________________________________________________________ 

 

Number of subjects = 77, number of observations = 77, number of failures = 77, 

time at risk = 2607, Wald chi-squared(5) = 10.6, Log likelihood = 8.38,  

Probability > chi-squared = 0.0594. 

______________________________________________________________________________________ 

 

a

The dependent variable or outcome is analytical time of the establishment of the park (“failure time” or “analysis 

time when record ends” —  thus, for the model, analysis time begins in 1950, and a science park that was 

established in 1983 has an analytical time of establishment of 33). The term “failure” refers to traditional 

applications of the survival-time model and the “survival” function, S.  As long as a “subject at risk” has not adopted 

the innovation by establishing a science park, it “survives” in the data, but on adoption it ceases to “survive” and 

leaves the set of potential adopters. t1 = aerospace/aeronautics; t4 = biotechnology/biomedical; the remaining 

technology categories (in the intercept here in Table 1) provided in AURRP (1997) are provided in the note to Table 

2 below where they are used. 

 

b

The z statistics and probability statements are for each of the underlying coefficients, rather than for the hazard 

ratios.