hossein hafezi; mahmod ekrami; nadergholi ghorchiyan; mohammadreza sarmadi
Abstract
The present study was conducted to fit the mathematical-structural model of knowledge commercialization in Payame Noor University. This research is quantitative and based on correlation method using Structural Equation Model (SEM) approach. Statistical population consisted of all faculty members and ...
Read More
The present study was conducted to fit the mathematical-structural model of knowledge commercialization in Payame Noor University. This research is quantitative and based on correlation method using Structural Equation Model (SEM) approach. Statistical population consisted of all faculty members and PhD students in Payame Noor University. In this research, the required data were collected by using the proportional stratified sampling method in two stages, pilot stage (100 participants including 79 faculty members and 21 PhD students) and main stage (245 participants including 200 faculty members and 51 PhD students). A research instrument comprising the researcher-made 114-item questionnaire designed to measure seven existing constructs in research model. Psychometric properties, including reliability and validity, measures of research instrument were evaluated and supported by using the confirmatory factor analysis approach on the data of preliminary stage. In order to analyses the data in the main stage, first by means of data screening (including box plot, Mahalonobis statistic, uni- and multivariate Skewness and Kurtosis coefficients, and scatter plot) the position of outliers' data and assumptions underlying Structural Equation Model (SEM) statistical approach were also investigated. Then, by using a statistical approach of structural equation model, the way the model was fitted and mathematical-structural relationships between its existing constructs were tested. The results showed that preliminary fitted model requires some reforms in format involving deleting of 8 direct insignificant paths among its constructs as well as adding paths of correlation variance errors between indicators of the model endogenous constructs. Thus, the final modified model was perfectly fitted.