本研究以美國核准通過太陽能光伏技術相關專利為樣本，期間從1995至2007年，以Probit model主要分析專利向後引證對於專利移轉機率的影響，發現專利向後引證對移轉機率有負向顯著效果，可能是向後引證越多該代表專利創新程度低或較不具有突破性之技術。之後逐步將專利向後引證數拆解，首先拆解為專利文獻引證與非專利文獻引證，發現專利文獻引證呈現正向相關，專利引證越多可能使該專利有較豐富的技術背景而使移轉機率提高，非專利文獻引證為負向相關，則是可能非專利文獻引證越多越屬較為基礎的技術，或是處於某項技術相對較不成熟期之專利，因此會減少移轉機率。專利文獻次數對於專利移轉有正向影響，而引用非專利文獻次數則是負向影響，在比較這兩項變數對專利移轉機率的影響，非專利文獻引證對移轉機率的負向影響程度，大於引用專利次數的正向影響，或許這也是造成專利向後引證對於專利移轉機率為負的原因。 最後再將專利文獻引證拆解成美國與非美國專利引證，非專利文獻引證拆解成科學引用與非科學引用，結果發現美國與非美國專利引證都是正向效果，顯示移轉不會受到引證專利國家的限制，也隱含技術外溢對移轉有正向相關。而科學引用與非科學引用皆為負向不顯著的效果，科學引用是指發表ISI-Web knowledge database的文獻，但科學引用呈現統計上的不顯著，而非科學引證是不包括發表在ISI-Web knowledge database的文獻、出版刊物或研討會成果，其素質無一評量標準，結果可能參差不齊，因此會減少專利移轉機率，另外，本文亦控制其他專利變數，發現與文獻大致上相同。In this research, we use the United States approved patent of the solar photovoltaic technology for the sample period from 1995-2007 to study the impact of the backward citation on the probability of patent transfer, With the Probit model analysis, we find that the patent backward citation has negative effect on the transfer probability. This also means the more backward citations, the lower the degree of patented innovations or breakthrough technique. Then we dismantle the patent backward citation step by step. We find that the patent literature citations are positively related to transfer probability because of the patent’s rich technical background. Non-patent literature citations have negative effect on transfer, because of the more non-patent literature citations mean the more basic on technology, or are relatively immature in a particular technical field Finally we disassemble patent literature citations into U.S. and non-patent citations, and also disassemble non-patent literature citations into scientific reference and non-scientific reference. The results show that both the US and non-US patent citations have the positive effects on patent transfer. This implies that transfer will not be restricted by the nation of cited patents, and there is a positive correlation between technology spillovers and transfer. Scientific reference and non-scientific reference are negatively affect transfer probability. Scientific references are those published literature of the ISI-Web of knowledge database, but scientific references are not statistically significant. Non-scientific references include conference proceedings, without an assessment standard, their quality may be difficult to evaluate, therefore reduce the probability of transfer.
