Probit and OLS regressions

The total sample is composed of 4,650 European patents. I run a Probit regression that estimates the probability of a patent being CL or DL, and two OLS regressions in which the number of inventors that take part in the development of a patent (NRINV) and the number of supplementary classes (NRCL) are the dependent variables. The regressors are coimtry, sectoral, regional and firms characteristics. Table 8 lists the variables used in the regressions. 

CL and DL Dependent variable of the Probit regression it takes the value 1 when the inventors are co-localised 0 otherwise 

NRINV Dependent variable of the OLS regression number of inventors that collaborate in a patent 

NRCL Dependent variable OLS regression number of supplementary classes in a 

LABS Regional density of chemical R D laboratories. This is given by the number 

---

The next section introduces the key issues of the chapter. Section 3 describes the geographical distribution of chemical innovations in Europe and presents some insights from the data about the effectiveness of the firm and the geographical cluster. Section 4 checks for multiple correlations by means of Probit and OLS regressions. The final section concludes. 

---