Discouraging Honorific Citation

The paper presents a study on analysis of citations made in the introduction section of articles that report empirical research. The attempt is to proceed with the validation that pertinence of citations in scientific articles is a core problem in citation analytics. Citations made in the introduction sections of research articles published in Nature weekly journal were studied. The entire research articles published from Volume 523 to 527 were studied. The result showed that the mean pertinence for the entire study is about 32%. Overall, about 68% of citations made in the introduction sections may not be applicable in the computation of effective impact of publications. It forms one of the first attempts to use empirical methods to determine pertinence of citations in scientific publications. Here in, the rationale for the study is identified


INTRODUCTION
The journal publication system has been identified to play a very important role in the science social system.Cronin 1 identified that the perceived role and importance of the system is considerable and includes ensuring the preservation of standards and screening of knowledge added to the literature.It is also believed that the publication process forms the basis for the allocation of scientists' rewards and recognition when their work is published or cited.Price,2 laid the foundation of the present reward system in the science social system, where citation is regarded as a means to distribute credits/recognition to the published scientists.
5][6] Specifically, the following were identified as problems of citation analytics: erroneous computation of citation based impacts; 4 Problematic results due to variable coverage of search engines as well as their availability in international bibliometric data bases. 7They can be directly or subtly gamed and manipulated by the editor through coercive citation or by the author through self citations. 8Adedayo, 9-11 also indicated that, because citations are not always generally positive, it is possible that the present usage of citation analytics, to implicitly allocate rewards and recognition to counterproductive efforts.Studies have been published to discourage honourific reward allocation. 6,9,12so, it has been proposed by Cawkell 13 that the citation analytics would work better, only if every citing author meticulously cited only the earlier works pertinent to theme of the new manuscript.Particularly, Adedayo, [14][15][16] identified that oftentimes, not all cited references express the same opinion with the manuscript where they are cited.If the issue of pertinence is adequately settled, then coercive and self citations would become of little or no significance.
In this particular article, an empirical investigation to study pertinent citations in the introduction sections of articles published in Nature, the weekly international journal of science published by Nature Publishing Group is presented.The idea presented in the report is very fresh, and original!It forms one of the first attempts to use empirical methods to determine pertinence of citations in scientific publications.Herein, the rationale for the study is identified The total number of authors cited in the Introduction sections were counted and recorded as Nc.Also, a counting of common citations made both in the Imaginary and the Real sections was made, and recorded as nc.Pertinence (p) of the Imaginary section (Introduction section) of each article was determined by finding the ratio nc: Nc expressed as a percentage i.e.
The average Pertinence, Nc, and nc for the journal publication in a particular volume was determined.The overall averages for the journal entire study were also determined.

RESULTS AND DISCUSSION
Tables 1 to 5 present the results for the study.Figure 1 presents the cumulative frequency distribution of pertinence for the study, while Figure 2 gives overview of pertinence distribution with frequency.By determining the upper quartile in Figure 1, it is clear that over 75% of the articles have pertinences below 50%.This shows that majority of citations made in the introduction section of the articles studied have not validly supported the reported study.In Figure 2, only articles with pertinences less than 65% have shown frequencies ≥ 2. Figure 3 gives the frequency distribution of Nc for the study.Here, it is shown that the most frequent Nc lies within the range 5-35.Within this range, Nc have mostly frequencies ≥ 1. Figure 4 presents the frequency distribution of nc within the articles analyzed.Frequencies were high for low nc values.These decreased down the line.
From here, it could be surmised that the probability of finding article with higher nc decreases.
Overall the average pertinence for the study is found by calculating the mean for the average pertinences for all the issues analyzed i.e.
. Where pm is the mean of the average pertinences for all the issues analyzed.With this result, it is clear that, on the average, only 32.6% of citations in the introduction sections of the articles studied are pertinent to the reported research.This result is supported by the predictions made by Adedayo. 10 In his study, Adedayo, 10 extended the work of Saha et al. 3 , drawing similarities between citations and votes.When citations are considered as votes, Adedayo 10 , predicted that majority of citations made in the introduction sections may not be applicable in the computation of effective impact of publications.

CONCLUSION
This study has shown that significant proportion of citations made in the introduction sections of scientific articles only have imagined pertinence to the study reported.The result of the study also supports the assertions of Adedayo,6 and Persson & Glanzel 16 which discouraged honorific attribution of rewards and recognition.The opinion of Cawkell 13 , that pertinence of cited literature reference in a scientific article is very important in impact evaluation considerations is also here by reinforced.The study, therefore, also suggests that citations in scientific articles can be validly classified into two i.e.Citations in Imaginary sections and citations in the Real sections.Pertinence; a new parameter useful in the evaluation of scientific publications has been introduced.

Figure 4 :
Figure 4: Frequency Distribution of n c within the Articles Analyzed

Figure 3 :
Figure 3: Frequency Distribution of N c for the Study

Methodology
Citation pattern in Research Articles published in Nature was studied.The weekly international journal, Nature, publishes different forms of articles such as: News & Views Articles, Research Article, Research Letters etc.Only citations in Research Articles were studied.Citation pattern in publication issues from Volume 523, Number 7560 to Volume 527, Number 7577 was studied.All Research Articles within this range were studied.A systematic cull of citation in the articles, as suggested and described by Adedayo 6,15,16 was adopted.Citations in the articles were classified as citations with Real and Imaginary Pertinence. 11,16Citations made in Introduction sections were considered as Citations with Imaginary Pertinence while those made in the Methodology/Results/Discussion of Result/Conclusions are considered to have Real Pertinence.Citations made in the Methodology/Results/ Discussion of Results/Conclusions are classified as Real Citations, because these truly show that the cited source support the new research being reported, and thus is pertinent to the reported study.Citations made in the Introduction sections are classified as Imaginary Citations.This because, any citation made in the Introduction that cannot be cited in the Methodology/Results/Discussion of Results/Conclusions can only be stated to have imagined pertinence to the study.The pertinence is only a figment in the imagination of the citing author.
Table 1 provides information on citation distribution of Introduction section in articles published in Volume 523 of Nature.From the table, the highest pertinence observed is 43%, with Nc of 42 and nc of 18.The lowest pertinences were 0%.The averages for the volume are pertinence of 32.5%, Nc of 21.2 and nc of 7. Table 2 provides the citation distribution in Volume 524.The highest pertinence observed is 100%, with Nc of 10 and nc of 10.The lowest pertinence is 23% with Nc of 22 and nc of 5.The averages for the volume are pertinence of 50.1%,Nc of 19.8 and nc of 9.6.Table 3 provides the distribution for Volume 525.The highest pertinence observed is 62%, with Nc of 21 and nc of 13.The lowest pertinence is 0%.The averages for the volume are pertinence of 30.7%, Nc of 18.7 and nc of 6.3.Table 4 shows citation analytics for Volume 526.The highest pertinence observed is 78%, with Nc of 9 and nc of 7. The lowest pertinence is 6%, with Nc of 34 and nc of 2. The averages for the volume are pertinence of 30.5%,Nc of 18.1 and nc of 5.In Table 5, we see the citation distribution for Volume 527.Highest pertinence observed is 47%, with Nc of 17 and nc of 8.The lowest pertinence is 9%, with Nc of 23 and nc of 2. The averages for the volume are pertinence of 19.45%, Nc of 28 and nc of 4.9.