Malaysian Journal of Library & Information Science, VoL1, no.2, December 1996: 89-93

LmRARY AND INFORMAnON SCIENCE LITERATUREAND LOTKA' S LAW

B. K. SeDCbe ADan bin Taib

Mohd Faris bin Hassan:rvfLIS Programme, F acuIty of Computer Scia'lce

& Infonnation Technology, University of Malaya50603 Kuala Lwnpur, Malaysia

ABSTRA CT

Attempts to test the validity of Lotka 's law in the domain of library and informationscience (LIS) taking the annual name index of Library and Information ScienceAbstracts (LISA) 1992, and annual author index of LISA 1993 as the base, whichincluded 7101 and 7591 abstracts of articles contributed respectively by 8284 and7664 authors. Lotka's law is found to be applicable with the value of n as 3.23 in the

first case and 3.1 in the second case. It is concluded that the value of n is found to behigher in LIS compared to exact sciences because the number of authors contributingtwo or more articles are less in this field

Keywords: Lotka's la~ Library and information science literature; Author productivity,Bibliometrics.

INTRODUCTION no excq>tion, where a very small numberof people hold major share of the wealth,whereas billions in the world are prac-tically destitutes.

In many instances in our life we find thatthere is a close link ba:weal less andmore. In the academic sphere we find thatfewer studmts pass with very high per-cmtage of marks. In the arma of sports itis Seal that fewer football players scoremore goals, fewer crick~ers score moreruns, fewer bowlers take more wickets,and so on. In the field of medicine, fewerdiseases take more lives. The rule appliesin the case of cities too - fewer cities are

mrtropolis or mega-cities harbouring mil-lions of people whereas smaller cities aremore in number with fewer number ofpeople. The field of economics is also of

As can be expected the rule applies to thesphere of library and infonnation scialceas well. It is but a common experialcethat the Ina."IQffium number of contribu-tions are accounted for by a few authors.In 1926, Alfred J. Lotka, a statistician ofthe Maropolitan Life Insurance Compa-ny, became algrossed with the idea ofdetennining, 'if possible, the part whicl1mal of differart calibre contribute to theprogress of scialce'. For this purpose, heused the index of Chemical Abstracts for

Sen, B.K. j Che .4:;lan B. T. and M ohd Faris B.H.

the years 1907-1916 and developed alisting of A and B names [i.e. the namesstarting with the la:ters A and B] and thecorresponding number of papers each au-thor produced. The same procedure wasapplied to Auerbach's Geschichtstajelnder PJl}'sik till the year 1900 using com-plete coverage (Lotka, 1926). The resultsobtained ri"DN surprising regularity whichallowed Lotka to derive the equation., x~= c where x stands for the number of

contributions, y for the number of au-thors, and c is constant (Lotka,1926).From these studies he found out the valueof n as 2. This finding finally becameknown as Lotka' s law or the inversesquare law of scimtific productivity. Thestudy stimulated a number of investiga-tions giVing rise to not only new findingsbut also controversies.

articles published in the Journal of Oil-seedS Research published bebNeen 1984and 1992 Kalyane and Sm (1995) folD1dthe value of n as 2.07. However, DeOliveira (1983) after analysing the litera-ture on jackfruit published bebNeen 1950and 1981 folD1d that the Lotka' s law didnot hold true in this case because thenwnber of researchers in the field wassmall. Conducting further studies DeOliveira (i984) realized that the appli-cation of Lotka's law would prove to berather difficult in the realm of socialscimces. Gupta's (1989) study with psy-chological literature of Africa for theperiod 1966 to1975 showed the validityof Lotka's law with a much higher valueofn as 2.8.

0 BJE CfIVE

Loth's study was related to the field ofscimce, where the number of contribu-tors in most fields is found to be high,and the rate of growth of the field interms of the number of items published ina y~r is gmerally found to be highcompared to other fields of knowlOOge.The studies conducted by others takingvarious fields of scimce also corrobo-rated Lotka's finding and the value of nwas found to be around 2. For example,Gupta(1987)foundthevalueofn as 1.9(for all authors), 1.8 (for the first author),2.2 (for single author), and 2.4 (for theco-author) whm analysing the Nigerianmtomological literature publishoo be-tween 1900 to 1973. While studying theauthor productivity in the field ofgeophysics, Gupta (1992) found the valueof n as 2.1. Analysing 498 research

Library and infonnation science is not anexact scia1ce like physics or chemistry.In this field the nwnber of rontributorsare less, and the growth of literature isalso not as high as it is found in manybranches of the exact sciences. Latkaconducted his study with the author in-dexes of Chemical Abstracts ~ years1907 to 1916. The volume of literatureproduced and the growth rate of thesubject during that period can be seenfrom Table 1. The growth during theperiod 1907 and 1915 was normal whichbecame stunted during 1915 to 1920because of World War I. Table 2 depictsthe scenario of library and informationscia1ce (LIS) literature as they appearedin LISA (Library and Information ScienceAbstracts) during the period 1984 to1993. Here we have takm the same soan

Libra')' and Information Scienc~ Lit~rature and Lotka's LaK,

1: Chemical Literature 1907-20 of years as did Latka. From Table 2 itappears that there was a slump during1985 to 1989. There is no reason tobelieve that the production of LIS litera-ture Walt down during those years. Thecoverage of articles from various pe-riodicals by USA in many cases is veryerratic (Sm, 1996) and this may haveresulted in the haphazard productivityscmario. The productivity of LIS litera-ture during 1~83 to 1992 is about one-half of that of the chemical literatureduring 1907-1916. Hmce, it was thoughtthat the authors' productivity might fol-low Lotka' s law and the study wasundertakm.

able 2: Library & Infonnation ScimceLiterature Baw~ 1984 - 1993

:MEffiODOLOGY

In the field of library and informationscimce, USA is found to be morecomprehmsive than other abstracting andindexing services in the field. Hmce, thisabstracting service was chosm for ourstudy. Till 1992 the name index of USAincl uded personal authors, corporateauthors, ac. For this Study only personalauthors were consideroo. The annualname index of 1992 and annual authorindex of 1993 were chosm for the studyas it includoo 8284 and 7664 personalauthors respect.ively. The number ofauthors contributing one, two, or morearticles each were countoo manually, andthe results tabulatoo ( Table 3A and 3B).

To find out ilie value of n. ilie studystarted wiili ilie premise of n = 2. The

values obtained were widely differmtfrom ilie real values (Tables 3A and 3B).As ilie calculated values were mum

Sen, B.K.; CheAzIanB.7: and i\lohd FarisB.H.

higher than the real values, the calcula-tions were carried out with the increasedvalues of n. In order to save time andshortm the procedure, the study deter-mined the value of n that matches withthe number of authors who have contri-buted two papers each using the follow-ing formula.

2n.771 = 7229~ 2n= 7229/771~ nlOg2 = log 9.376~ n(0.301) = 0.972~ n = 0.972/0.301

~n=3.23

Using the value of n = 3.23, the number

of authors contributoo three, four, or fivearticles each were computed (Table 3A).Similarly, ~e same procedure was adop-too for the 1993 data and the value of nwas found to be 3.00. With the value ofn= 3, the calculatoo values of authors

contributing three or more articles werefound to be differmt from the observedvalues. However, \\ith the value of n =

3.1, the observed and calculated valueswere found to be very close (Table 3B).

x~ = c. (eqn. 1)

Putting the value of x = 1, and y = 7229,

( videTable 3A), the calculation obtainedwas',

1°.7229 = C=>7229 = C

Putting the value ofx = 2, and y = 771,

Table 3A: Author Productivity based on LIS..4 1992 Data

Table 3B: Author Productivity based on LISA 1993 Data

and C = 7229, the calculation obtained

was;

Library and lnfonnation Science Literature and Lotka's Law

produCtiVity trends. Annals of Libra-ry Science and Documentation, 39no.4: 145-52.

RESULTS AND CONCLUSION

Tables 3A and 3B indicated that thenumber of authors obtained with the va-lue of n = 2 is widely differmt from the

real values. However, with the value of n= 3.23 in the first case and n = 3.1 in the

second case, the calculated values arefound to be very close to the real values.Hmce, the study concludes that Lotka' slaw is applicable in the field of libraryand infonnation scimce with mIJdl highervalues whm compared to the exact scim-ces. 11ris is because the number of au-thors contributing 2 or more articles areless in this particular field compared to ascimtific field where the number of suchauthors would have approximated thefigures giVal in column 3 in Tables 3Aand 3B.

Kalyane, V. L. and B.K.Sen. 1995. A bi-blioma:ric study of the Journal of Oil-seeds Res earch. A nnals of Li b raryScience and Documentation, 42 no.4:121-41.

De Oliveira, S. M., et at. 1983. Aplica-cao da lei de. produktividade deauto-res de Latka a literatura da jaca. Re-view Biblioteconom Brasil, 11 no.l0:125-30.

De Oliviera, S. M. 1984. A lei de Latkasobre a productividade autres; appli-calidade do quadrado inverso. ReviewEscola Biblioteconom UFlvKl, 13 001:207-33.

REFERENCES Gupta, D. K. 1989. Lotka' s law and itsapplication to author productivity dis-tribution in psychological literature ofAfrica for the period 1~1975. Hem/dLibrary Science, 28 no. 1-2: 11-21.

Latka, A. J. 1926. Statistics - the frequ-

ency distribution of scimtific produc-tivity. Journal of the WashingtonAcademy of Science, 16: 317- 25.

Sat, B. K. 1996. Contribution of AsianUS journals towards the developmentof US profession in Asia during thepast decade as viewed through the bi-bliometric lms In: Congress of South-east Asian Librarians, 10th, 1996.Kuala Lumpur: CONSAL X, Vol.l:195-220.

Gupta, D. K. 1987. Latka's law and pro-ductivity pattern of mtomological re-search in Nigeria for the period 1900-1973. Scientometrics, 1291 no.2: 33-46.

Gupta, D. K. 1992. Scimtometric studyof exploration geophysics - author

.I

93