Predicting technology transfer from data in research on technology development. (25th February 2005)

• Record Type: Journal Article
• Authors: Chambers, David W.; Leknius, Casimir; Reid, Laura
• Is Part Of: International journal of technology transfer and commercialisation. Volume 4:Number 3(2005)
• Page Start: 318
• Page End: 340
• Publication Date: 2005-02-25
• Subjects: technology transfer -- generalisability analysis -- robust technology -- dentistry -- variance analysis -- predictability -- technology adoption; Technology transfer -- Periodicals; 338.926
• Languages: English
• ISSNs: 1470-6075

Abstract

Neither the average effectiveness nor the variations of technology in transfer applications can be adequately estimated directly from performance data in standard research studies designed to develop technology or to "prove" its suitability for markets. The characteristics of the adopter and the situation to which the technology is transferred affect the estimates of outcomes most likely to be observed. A statistical procedure – generalisability analysis – is presented that can be used with various development or qualifying research to estimate variance associated with various factors that might influence technology transfer. By combining these variance estimates with known characteristics of specific applications, adopters of technology can estimate likely effects and their variance when applying technology. The use of generalisability analysis to estimate components of variance in technology transfer is illustrated with data from a study involving the fabrication of temporary crowns in dentistry. A three-factorial, fully crossed, random effects model was tested with a range of user skill values and two values on a technology and situational factors (U × T × S). It is shown in this case that the user contributes the largest amount of expected variance in normal transfer applications, followed by the interaction between the user and technology, complex interactions and error, and technology. It is demonstrated how regressions analysis and confidence intervals partitioned by expected sources of variance can be used to estimate the most likely results of technology transfer to various application contexts based on data provided in technology development or qualification studies. General observations about the nature of technology transfer are also offered.