Measuring a Model on Credibility Evaluation of Scientific Websites: Exploring Relationships and Priorities

Abstract Owing to the extreme importance of evaluating the credibility of existing scientific websites, the present study sets out to measure a proposed model concerning the views and preferences of university students in Iran, when evaluating information. Data were collected by administrating a highly validated questionnaire among 487 students in ten top universities in Iran. Structural Equating Modelling using software SmartPLS was conducted to analyse the data collected. To evaluate the measurement and structural models, a set of criteria including Cronbach's alpha, factor loadings, convergent and divergent validities, R2 , Q2 , redundancy, and GoF were considered to measure the power and validity of the model. Considering the path coefficients and the t-statistic for the dimensions and their components, path analysis showed that the t-statistic is greater than 2.57 indicating that all of the constructs contributed to the credibility of information on scientific websites at 99% confidence level. Coefficients of correlation concerning the overall information credibility were found to be 0.728 for trustworthiness and 0.718 for expertise. Expertise with a path coefficient of 0.968 and trustworthiness with 0.948 were the first and second priorities for the main variable. Moreover, ethics with a path coefficient of 0.787 and objectivity with 0.464 were the first and last priorities for trustworthiness respectively while accuracy with 0.874 and professional information with 0.674 were the first and last priorities for the dimension expertise respectively. The model could be used to evaluate the credibility of scientific websites’ information, which also provides a potential set for further research.