Nobel Prize winners 2016: Igniting or sparking foundational publications?

A warning for Chinese academic evaluation systems: short-term bibliometric measures misjudge the value of pioneering contributions

“…Indeed, in previous investigations Rousseau, 2016, 2017) we found that most fundamental work of Nobel Prize winners, such as the papers of Jean-Pierre Sauvage, J. Fraser Stoddart, and Ben L. Feringa (Nobel Prize winners in Chemistry in the year 2016) have the same "sparking" characteristics. Concretely, 23 of 29 publications reported by RSAS (2016b) have been shown to be influential in this way (Hu and Rousseau, 2017). Ruska's fundamental work (Ruska, 1933) on electron optics entitled "The electronelectron microscopic image of irradiated Surfaces" (a key article that let him to receive a Nobel Prize in Physics in 1986; https://www.nobelprize.org), behaved similarly.…”

Do citation chimeras exist? The case of under‐cited influential articles suffering delayed recognition

“…Stimulated by recent research by ourselves—about under‐cited influential publications (see further) (Hu & Rousseau, ; Hu & Rousseau, ) and by others about delayed recognition (Du & Wu, ; Ke, Ferrara, Radicchi, & Flammini, ; Li & Shi, ; Li & Ye, ; van Raan, , ) we became interested in the question if an article can suffer delayed recognition (an aspect of its first forward citation generation) and at the same time be under‐cited influential (an aspect mainly determined by subsequent citation generations). In the natural sciences rare events are often of extreme importance.…”

“…All the types of articles mentioned above are based on the—direct—citation distribution. Yet, recently Hu and Rousseau (, ) introduced another type of articles in which first, second and third citation generations play a role. Such articles, referred to as under‐cited influential articles, are characterized by three properties: (a) they are reasonably well‐cited (a basic requirement to be influential); (b) citations of citations (second generation citations) are rather high, so that the original one is influential in an indirect way (a more refined token of influence); (c) given condition two, these articles received fewer citations than expected (being under‐cited).…”

“…Such articles, referred to as under‐cited influential articles, are characterized by three properties: (a) they are reasonably well‐cited (a basic requirement to be influential); (b) citations of citations (second generation citations) are rather high, so that the original one is influential in an indirect way (a more refined token of influence); (c) given condition two, these articles received fewer citations than expected (being under‐cited). Each of these three requirements is operationalized leading to two formulae referred to as sparking indices (Hu & Rousseau, ) and explained in more detail further on. Although the framework we use to describe our results is in principle database neutral (admitting that some databases are more suited to find the data we need, for example, Clarivate Analytics' Web of Science (WoS) is much easier to use than Google Scholar).…”

“…Given a publication A, its sparking index derived from the top 1% citations, and denoted as S 1 (A), is defined as (Hu & Rousseau, ): where α 1 is the median value of the citations received by the top 1% most cited articles citing article A and α 2 is the median value of median values of the number of citations received by articles in the previous top 1%. The first median reflects the reaction in the second citation generation, while the second median reflects the reaction in the third citation generation of article A.…”