29 June 2008

Message

Message

Message

Message

Message

Message

Statistics Day 2008

2

Blank

3

Statistics Day 2008

PROF. P.V. SUKHATME

(27 July, 1911 – 28 January, 1997)

Pandurang Vasudeo Sukhatme was bornto Vasudeo Hari Sukhatme and SatyabhamaSukhatme on 27th July, 1911 in the villageBudh, district Satara, 100 miles south of Pune.After completing his school education in Punehe graduated in 1932 from Fergusson Collegewith Mathematics as the principal subject andPhysics as a subsidiary subject. During 1933-36, he studied at the University College,London and was awarded a Ph.D in 1936 anda D.Sc. Degree in 1939 for his work onbipartition functions. This work waspublished in the “Philosophical Transactionsof the Royal Society of London, Series A”, June,1938.

Whilst in London, Prof. Sukhatme cameunder the influence of such eminentauthorities in Statistics as R.A. Fisher, JerzyNeyman and E.S. Pearson and did valuableresearch in Statistical Theory of Sampling, histwo most significant contribution being, oneto bipartitional functions under the guidanceof R.A. fisher and the other to sampling theoryentitled “Contributions to the Theory of theRepresentative Method” under the guidanceof J. Neyman and E.S. Pearson. The latterpaper laid solid foundations for his subsequentpioneering research in the sampling theory ofsurveys and improvement of agricultural

statistics which ushered in what may beappropriately termed as the Sukhatme era inthe development of agricultural statistics inIndia and the world.

When P.V. Sukhatme returned to India,while searching for a University job, he hadan interview with the Late Pandit MadanMohal Malviya, Vice Chancellor, BanarasHindu University. Though Panditji wassatisfied with the brilliant career of P.V.Sukhatme and agreed to create a Departmentof Statistics in the University to accommodatehim but he wanted to know from Sukhatmehow a Chair in statistics would help our poorcountry - India. P.V. Sukhatme did not knowhow to answer this question nor did he joinBanaras Hindu University but this questionmust have moved him sufficiently indetermining his future file, particularly hisfundamental work on nutrition.

During 1939-40, he was a Professor at theAll India Institute of Hygiene and PublicHealth, Calcutta. In 1940 he joined ICAR as aStatistician, and was later on appointed asStatistical Advisor to the Council to head itsStatistical Unit. On account of his dynamicleadership, following the path and tradition setby him, the statistical branch of ICAR

Statistics Day 2008

4

eventually grew to become a full-fledgedInstitute (Indian Agricultural StatisticsResearch Institute) exclusively devoted toresearch in Agricultural Statistics. In thecontext of the green revolution the importanceof statistical techniques in agricultural researchhardly needs any emphasis.

Prof. Sukhatme, as a founder of the IndianSociety of Agricultural Statistics, devoted agood deal of his time and energy to thepopularization of statistical methods amongthe practitioners of agricultural, veterinary andrelated sciences. He served as the FirstHonorary Secretary of the Indian Society ofAgricultural Statistics for a number of years.The Society owes him a lot for his continuedvaluable guidance as well as for shoulderingthe responsibility, as its President during 1991and Executive President since 1970 till hisdemise. From its inception to 1963 he workedclosely with Dr. Rajendra Prasad (FounderPresident of the Society), the then Minister forFood and Agriculture and later President ofIndia.

In 1951, he was a Visiting Professor at IowaState University, Ames Lowa, USA where hecompleted his textbook on sampling. During1952-70, he headed the Statistics Division ofthe Food & Agriculture Organisation(FAO) ofthe United Nations in Rome. After retiringfrom the UN in 1971 he served as RegentsProfessor, University of California at Berkelyand then settled in Pune, carrying out valuablework on nutrition at the MaharashtraAssociation for the Cultivation of Science. Heauthored several books on the variousscientific topics of interest and published morethan 200 research papers in reputed nationaland international journals.

Prof. Sukhatme was well known in the fieldof nutrition for the Sukhatme-Margen

hypothesis which in plain language implies thefollowing: At low levels of calorie intake,energy is used with greater metabolicefficiency and efficiency decreases as the intakeincreases over the homeostatic range.

He was awarded the Guy Medal by theRoyal Statistical Society for his paper onnutrition which he presented to the Society in1963, the B.C. Guha Memorial Lectureship ofthe Indian Science Congress Association in1965 and the B.D. Tilak Lectureship of theIndian National Science Academy in 1982.Among the numerous other honours he hadreceived, mention must be made of theFellowship of the American StatisticalAssociation, National Academy of Sciences,Allahabad, Indian Academy of Sciences,Bangalore and Indian National ScienceAcademy, New Delhi. He was electedmember of the International StatisticalInstitute, Netherlands and its Vice Presidentin 1969-70. For his outstanding contributionto Science and Human Welfare, he wasconferred the Padma Bhushan by the Presidentof India in 1973. He was awarded the HariOm Ashram Trust Award by the UniversityGrants Commission in 1983. For thedistinguished service to the cause of Statisticsand its application to agriculture and alliedfields, he was conferred with the honour ofSankhyiki Bhushan in 1989 by the IndianSociety of Agricultural Statistics, New Delhi.He also received the P.C. Mahalanobis Awardat the Jaipur Session of the Indian ScienceCongress Association.

Prof. Sukhatme expired on 28 January,1997 at Pune. With the demise of Prof.Sukhatme, the scientific community in generaland the Agricultural Scientists in particularhave lost a great statistician, true advisor,dynamic leader, well-wisher of humanity anda renowned personality of international fame.

5

Statistics Day 2008

PROF. B. L. S. PRAKASA RAO

JAWAHARLAL NEHRU CHAIR

PROFESSOR

DEPARTMENT OF MATHEMATICS

AND STATISTICS

UNIVERSITY OF HYDERABAD

Prof. Bhagavatula Lakshmi Surya PrakasaRao, was born on 6 th October 1942 atPorumamilla, Andhra Pradesh. He did B.A.(Hons.) from Andhra University in 1960.Having done M. Stat. from Indian StatisticalInstitute, Calcutta in 1962, he obtained Ph. Ddegree from Michigan State University in 1966.Currently he is serving as Jawaharlal NehruChair Professor, Department of Mathematicsand Statistics, University of Hyderabad,Hyderabad.

He studied the asymptotic distribution ofthe maximum likelihood estimator in non-regular cases such as estimation of the locationof the cusp for a family of densities whichincludes double exponential; developed cuberoot asymptotics underlying the basicdistribution theory for the study of themaximum likelihood estimator of a unimodaldensity under restrictions on the shape ofthe density; contributed extensively to theparametric and nonparametric inference forstochastic processes and characterizationproblems and limit theorems in probability.

During the last 40 years, Prof. Prakasa Raohas been working on the study of asymptoticinference of parametric and non-parametricestimation for continuous as well as discrete

time stochastic processes. He has developedthe basic asymptotic theory for the estimationof the drift parameters of a diffusion process.Prof. Rao’s recent work consists of estimationof parameters for processes driven by afractional Brownian Motion which is veryimportant for modeling problems connectedwith traffic patterns of packet flows in highspeed data networks such as the internet.

Having published over 200 papers innational and international journals, he hasreviewed over 60 books for international andnational journals and has authored,coauthored and edited ten books dealing withparametric and non-parametric inference forstochastic processes. He has guided studentstowards their Ph. D on inference for stochasticprocesses. He has been editor or a member ofthe Editorial Board of international andnational journals such as Statistical Inferencefor Stochastic Processes, Statistics andDecisions, Statistics and Probability Letters,Electronic Journal for History of Probabilityand Statistics, Sankhya and Indian Journal ofPure and Applied Mathematics.

He was the Director of Indian StatisticalInstitute during 1992-95. He was also aChairman or a Member of various committees

Statistics Day 2008

6

appointed by CSIR, Government of India andwas editor of publications, Indian NationalScience Academy during 1991-93.

He has served as Visiting Professor (UPE)2004-2006, Distinguished Scientist andDirector, Indian Statistical Institute, Kolkataduring 1992-1995, Professor, Indian StatisticalInstitute, Delhi during 1976-1986 and 1988-1991 and Distinguished Scientist, IndianStatistical Institute, Delhi during 1992-2004.He has also served as Assistant Professor andAssociate Professor at Indian Institute ofTechnology, Kanpur. He has been a VisitingProfessor at University Of Iowa, Iowa City,USA; University of California, Davis, USA;University of Wisconsin, Madison, USA;University of Illinois, Urbana, USA.

He was awarded gold medal and a specialcertificate of merit by Andhra University forrecord breaking performance at the B.A. Hons.

in 1960. He received Bhatnagar Award forMathematical Sciences (1982). He is an electedmember of the International Statistical Institute(1982). He was elected Fellow of the Instituteof Mathematical Statistics, USA (1983); electedFellow of the Indian National ScienceAcademy (1984); elected Fellow of the IndianAcademy of Sciences (1992); elected Fellowof the National Academy of Sciences (1993);elected President of the Section on Statistics(83rd session ) of the Indian Science CongressAssociation (1995-96). He was selected asNational Lecturer, University GrantsCommission (1985-86). He was awardedoutstanding Alumni Award by Michigan StateUniversity 1996. He was also awardeddistinguished service award by MathematicalAssociation of India 1999. He was awardedbest paper in the field of StatisticalMethodology by the Indian Society ofAgricultural Statistics 2000.

Statistics Day 2008

8

Blank

9

Statistics Day 2008

PROFESSIONAL TRAINING IN STATISTICS

by P. C. Mahalanobis1

The demand for trained statisticians issteadily increasing in India, but the statisticaleducation given in the universities has notbeen found adequate for professional work. Itis recognized that there is a great need ofproviding professional training of a highquality for a statistical career.

The nature of professional work in statisticsvaries widely, and it is convenient to break itup into 4 or 5 special fields (each of which, ofcourse, may be further sub-divided into morespecialized sectors).

(i) Official and administrative statistics inGovernment and semi GovernmentDepartments or on the management sideof large commercial and industrialconcerns.

(ii) Statistical work in connection witheconomic analysis and planning inGovernment, semi-Government, orcommercial and industrial offices.

(iii) Applications of mathematical or analyticstatistics in scientific and technologicalwork and research.

(iv) Statistical sampling in various kinds ofsurveys.

(v) Statistical education, training andtheoretical research.

A comprehensive programme forprofessional training must include threethings. First, training in certain statistical“crafts” or “skills” which would be same forall five groups; secondly, a basic educationalcourse also same for all groups; and, thirdly,specialized courses for each group separately.

The first group, official and administrativestatistics, offers the largest volume ofemployment in India, and although a largenumber of posts are being created and filled,the quality of work is still low for lack oftrained personnel.

The demand for specialists in economicanalysis and planning has just begun and theshortage of competent statisticians is mostacute in this sector. In the third sector (that is,in the use of mathematical or analytic statisticsin scientific and technological work) theposition in India is appreciably better. Presentarrangements for education and training aredirected mainly or almost exclusively to thisparticular group.

(In consequence little or no attention has beengiven to statistical craftsmanship with the resultthat statisticians with a great deal of knowledgeof mathematical statistics have often been unableto deal with the professional work in official andadministrative statistics).

There has been a rapid growth in the useof statistical sampling in India, and the supplyof sampling statisticians falls far short of thedemand because facilities for training in thissubject are very meager.

Training and opportunities for research inmathematical and analytic statistics have,however, expanded in India during the last 10or 15 years and the position is notunsatisfactory. The supply of teachers for theusual academic type of statistical education hasalso expanded fairly rapidly.

1Bulletin of the International Statistical Institute Vol 33 No 5 (1951)

Statistics Day 2008

10

But basic training in what may be calledthe “statistical crafts” have been neglected withunsatisfactory consequences. I shall considerthese “crafts” under a number of heads.

Preparation of questionnaires andschedules: Training and skill in preparingquestionnaires and schedules are essentialrequirements for professional work.Unfortunately, in India, this subject is almostcompletely neglected. I have continually founda kind of feeling among mathematicalstatisticians that the framing of a questionnaireor a schedule is not their concern. Inconsequence, questionnaires and schedules,which have been used in the past or are stillcurrent, are often of a poor quality. Usuallythere is much waste of paper together with lackof space needed for entering the data in alogical manner. The arrangement is oftenhaphazard and without any reference to theprocedure to be adopted in collecting thematerial. Concepts and definitions are notproperly worked out, and so on. My ownexperience has been that it takes a good dealof time and practice for a junior statistician tolearn how to prepare good forms andschedules

Collection of primary data: It is necessary,in my opinion, for every professionalstatistician to have some personal knowledgeand experience of actual methods of collectionof data. Unfortunately this important aspectof statistical work usually has no place in theteaching programme. Most of the statisticiansworking in Government offices have,therefore, no idea of the difficulties or theconditions under which the data are actuallycollected

Scrutiny of primary data: R. A. Fisher hasstated somewhere that the first duty of astatistician is to cross-examine his data. In Indiaso far very little attention has been given tothis important subject. One reason is that, as

“STATISTICS MUST HAVE A PURPOSE”

…..Before starting to collect any new statisticsit is useful to pause and enquire “for whatpurpose?” When a statistician is requested to collectsome statistics his first responsibility is to ask theperson or authority making the request to explainas clearly as possible the purposes for which theinformation would be used. One may find, forexample, that statistical data are already availablewhich would serve the purpose. Or, it may turn outthat information of some other kind than thatsuggested at first would be more suitable. Also,when a complete census is suggested, a samplesurvey would be often found fully adequate……..When the National Sample Survey first startedcollecting information on unemployment in Indiathe concepts and definitions adopted were broadlythe same as those used in the western countries. Itwas, however, soon found that the western conceptof unemployment was inadequate because millionsof persons in India work on their own account inhousehold enterprises. They may have to sit partlyidle for lack of work; but they can neither gain norlose paid jobs. It became necessary, therefore, tothink not only of paid jobs, which one may get orlose but also of the amount of work one is doing, orthe amount of money he is earning by such work. Itwas necessary to introduce the concept of “normalperiod of work” and “normal earnings”; and tocollect information on the actual period of timeduring which the person is gainfully engaged inwork and express this period” of time as a fractionof the normal period of work; or to ascertain theactual amount earned in a given period and expressthis amount as a fraction of normal earnings; andto use these .fractions as measures of employment.Sometimes it is not possible to speak of the degreeof employment or unemployment (even in termsof period of work or of earnings) of a singleindividual; and it is necessary to refer to thehousehold enterprise as a. whole. It is necessarycontinually to construct new concepts anddefinitions to suit the needs of underdeveloped

countries. ##

Presidential Address: Third Pakistan StatisticalConference, Lahore, 1956 mentioned above,statistical developments in India have beenheavily concentrated in the third sector inwhich the statistician is usually associated withexperienced scientific workers who undertake

11

Statistics Day 2008

the responsibility for making the scientificobservations or collecting the statistical data.For example, in meteorology a great deal ofattention is given, as a routine measure, to thecontinuing calibration of meteorologicalinstruments. In agricultural and biologicalexperiments the scientific worker is muchconcerned with controls and the validity of theobservational data. In consequence, at least inIndia, it has not been necessary for thestatistician to give attention to the accuracy ofthe data and he has been free to use his analyticmethods without much primary scrutiny.Because of the collaboration of other scientificworkers who have undertaken theresponsibility of producing reliable data, thisarrangement has worked satisfactorily in thecase of meteorological observations andcontrolled experiments in the laboratory or thefield.

Lack of training and experience in cross-examining data has had however mostunsatisfactory consequences in the field ofeconomic and social statistics, and also in large-scale surveys where the data have to becollected by ordinary investigators withoutscientific training, who are usually scatteredover large areas and over whom the statisticianhas little or practically no control In official andadministrative statistics the result has been thegrowth of a “ritualistic” attitude of mind (toborrow the phrase aptly used by Stuart A.Rice). Most of the official and administrativestatistics which comes to Government of Indiais collected through agencies at the peripheryover which the Central Government often haveno direct control. The statistician who dealswith the material in the Central Governmentusually has no knowledge about the methodsof collection of the primary material or theagency or men employed for this purpose. Hesimply accepts the data as something given.He has no curiosity or desire to question the

accuracy (except in the very formal or trivialsense of reconciling misprints anddiscrepancies in totals or the results of similararithmetic calculations). A senior statisticianin Delhi told me, for example, that he did notknow of a single instance in which the validityof a figure was ever challenged in his wholeexperience.

A second consequence has been the growthof a somewhat legal or formalistic attitude. Afigure is received from State Government orother agency, and it is usually assumed thatthe Centre has no right (or in any case, it is notproper) to challenge its validity. It is somethinglike the record of a first or original Court ofLaw, which has come up to a Court of Appealwhere the evidence cannot be re-opened butarguments can be made only on points of law.(It is gratifying to note, however, that thepresent Census Commissioner is conducting,for the first time in India, a check on the censusenumeration).

What is true of the Central Government isalso generally true of the State Governmentbecause here also the statisticians usually dealwith data, which come up to them fromsources and agencies at the periphery. The

Smt. Indira Gandhi, Prime Minister of India, inaugurating the Delhi

Campus (subsequently a centre) of ISI, in the presence of P.N. Haksar,

Chairman of the ISI council, on 31 December 1974. Prof. B.S. Minhas

was also present.

Statistics Day 2008

12

State statistician also deals with the materialas something given. In this way a great deal ofstatistics is collected and sent up to the CentralGovernment the reliability of which is notquestioned or examined at any stage.

Because of this legalistic or ritualisticattitude towards primary data, statisticiansthemselves have sometimes tended to take apurely formal view of statistical work. Thestatistician puts up some statistical data foranalysis, and finds that this is not to the likingof his immediate superior officer or somebodyhigher up in the hierarchy (possibly even theMinister himself) who desires to have someother kind of analysis or even data which inhis intuitive judgment is more realistic. Usually(but, of course, not always,) the statisticianfeels that, as a good civil servant, it is his dutyto change his analysis or even his data.

In sample surveys my greatest difficulty inIndia has been to make highly trainedstatisticians (with a great deal of knowledgeof mathematical statistics) realize theimportance of investigator bias and responseerror. All their attention is focussed on thesampling error. In fact, the general attitude isto look upon the non-sampling error assomething, which does not concern thestatistician, or in any case is a kind of dirty job,which a highbrow statistician need not botherabout. This is why the use of inter-penetratingsamples has been persistently opposed in Indiaas sheer waste of money.

I have dealt at some length with thisparticular point because in my opinion the lackof interest and skill in cross-examining theprimary data constitutes the biggest singlesource of weakness of Indian statistics at thepresent time, and makes a great deal of thework in social and economic statistics more orless formal and unreal.

Computational work: A certain amount oflaboratory practice in computation work isincluded in many (but not all) statisticalcourses in India. Sufficient emphasis is,however, not always given. One of my greatestdifficulties has been to persuade a young M.Sc.in statistics to take up himself any seriouscomputational work. Statistical education (andtraining) in India is often somewhat formal sothat the student tends to think that his realconcern is with mathematical and algebraicsymbols, and that computations andcalculations can be done by clerks (who belongto an entirely lower order of beings). I haveoften quoted (but usually without much effect)R. A. Fisher’s statement that he learnt all hisstatistics through computations. In India, thisnegative attitude towards computational workis also partly conditioned by the generalresistance against manual work, which isconsidered to be derogatory to the dignity ofeducated or professional workers.

Organization of statistical processing:Even when the statistician has done a gooddeal of computational work in the practicalclass, he usually has no experience of theorganization of processing of statistical workby computers. He has no knowledge of properprogramming or of making even roughestimates of the time likely to be required tocomplete a job. In consequence, trainedstatisticians with high academic or researchqualifications are often unable to cope withlarge scale processing work. My impression isthat there is great deal of inefficiency and wasteof money in such work in India.

Costing and budgeting: Because of the lackof knowledge about time requirements, thestatistician is often unable to prepare realisticestimates of cost. I have seen even senior

13

Statistics Day 2008

statisticians in India giving estimates of costsometimes less than one-fourth or one-fifth ofthat actually required in practice or sometimeseven four or five time higher.One seriousconsequence has been the lack of reality in thescrutiny of estimates of statistical projects bythe Finance Ministry or Finance Departmentsof State Governments. Sometimes the budgetis cut down in such a way as to seriouslyhamper the work. Sometimes much largergrants are sanctioned than are really necessary.Estimates are prepared in a more or lessstandard administrative pattern. For example,if there is one senior statistician it wouldautomatically follow that he would have somany assistants, so many research workers, somany investigators, etc., irrespective of thenature of the work. This is quite natural asestimates tend to be judged by yardsticks,which have sole reference to work of a moreor less routine type. Also, quite naturally, thereis a tendency to put up much higher estimatesthan are considered really necessary on theview that the Finance Department would cutit down in any case by 50% or some suchmargin. This has led to a great deal of waste in

both men and money at least in theGovernment of India.

Preparation of statistical tables andstatements: Like the preparation ofquestionnaires and forms, the preparation oftables and statements does not find any placein most teaching programmes, and inconsequence tends to be looked upon as noconcern of the highbrow mathematicalstatistician. Usually tables and statements inGovernment offices are prepared by juniorclerks, and often hide those very aspects, whichrequire to be brought out. And yet the tablesand statements are the final end products onwhich depend to a large extent the impressioncreated on the mind of the administrator. Myexperience has been that it takes a great dealof time and practice to acquire skill in thepresentation of statistical data.

Preparation of statistical reports: Thestatistical report contains the final results of along process of statistical work of variouskinds. If the statistics are to be used in any way,such use must be based on the statistical report.It is necessary and desirable that a great dealof attention should be given to this part of theprofessional work. Unfortunately, usually thisagain has no place in the teaching programme.I have seen many statistical reports, which arelacking in organization of thought, logicalorder of presentation of facts and conclusions,and written in a slovenly style. When care isgiven to the writing of the report, the tendencysometimes is to make it a kind of imitation ofa technical paper suitable for publication in ascientific journal. Technical terms and jargonsare used in abundance, which simply annoythe administrator and create resistance. On theother hand, sometimes the reports are writtenin a highly polished administrative style, inwhich the entire attention is given to the form

A. Kosygin, first deputy chairman, USSR Council of Ministers, with

Pandit Jawaharlal Nehru and P.C. Mahalanobis 1961

Statistics Day 2008

14

and the surface phrasing without bringing outthe significant statistical facts or conclusions.A third type of report, which, also isunfortunately too common, deals in a veryelaborate and theoretical manner with smallchanges in say a statistical series which maybe entirely due to chance fluctuations orsimply to the unreliability of the primarymaterial. Economists without professionaltraining in statistical are often inclined tohandle the statistical material in this superficialway which has created a good deal ofunfavorable impression in administrativecircles.

I have drawn attention to eight handicraftsof statistics which are much neglected inteaching programmes in India. Basic trainingin these subjects should be given to all studentswho intend to take up statistics as aprofessional career. These statistical crafts areof particular importance for work in officialand administrative statistics. Professionalworkers desiring to enter these sectors should,in my opinion, have a great deal of advancedtraining and practice (besides the basic course)in these subjects.

Basic training in these subjects wouldimprove the quality of work of statisticiansengaged in the application of mathematicalstatistics in science and technology. Suchtraining should also be obligatory for thosewho wish to take up statistical education andresearch as their profession. I am prepared toagree, however, that such training is notnecessary for a person capable of doing“intuitive” work at the highest level. If, forexample, a Ramanujan is discovered in thefield of mathematical statistics he may be givenfull freedom to work in his own way.

Much thought has been given during thelast 4 or 5 years in many countries of the world

to the basic educational course dealing withthe general theory and practice of statisticalmethods. A good deal of this basic course canbe conveniently and profitably integrated withgeneral education. I have discussed in aseparate paper (recently prepared for aUNESCO publication) the possibility ofintroducing mathematical statistics as a subjectof study in secondary schools.

RESEARCH AND TRAINING

“…… We are now trying to provide professionaltraining in the Indian Statistical Institute, the CentralStatistical Organization or within the NationalSample Survey and other agencies. Here also I feel itmy duty to throw out a warning. There is a tendencyin certain quarters in India (and in other countriesalso) to try to make out as if “mathematical statistics”is a subject by itself without any reference to anyapplications to scientific research or practicalproblems. I must confess I do not understand thisposition. To me statistics must always have a purposewhich may be, of course, to help in drawing validconclusions from scientific observations for purelytheoretical researches (for example, the distributionof galaxies of stars or other problems in astrophysics);or, to help in selecting best varieties for agriculturalproduction; or, to help in industrial or economicdevelopment. There is plenty of scope formathematical researches in connection with statisticalproblems. I should like to give all possibleencouragement to really able mathematicians to enterthe field of statistics. But I should still hold thatstatistics is not a branch of mathematics but a fieldfor the application of mathematics. Pure mathematicshas its own justification independently ofapplications. But immediately one thinks ofapplications there must be a reference to somepurpose that is, to some contingent problem requiredto be solved. One must then work on contingentrelations, which are capable of being observed ormeasured and on the basis of which it would bepossible either to draw valid inferences in scientificmatters, or to take decisions in practicalproblems…..” ##

## Presidential Address: Third Pakistan Statistical

Conference Lahore, 1956

15

Statistics Day 2008

(b) to analyze, integrate and make a criticalappreciation of relevant information ifalready available; and

(c) when such information is not available,to prepare a project for the collection ofthe required information and theirsubsequent analysis and integration; also

(d) when a certain policy or actionprogramme has been decided to developstatistical controls for the assessment ofthe progress of implementation; and

(e) to recommend changes in the policy oraction programmes in the light ofexperience based on statisticalinformation of the required degree ofvalidity and precision.

These are the higher skills or techniques,which a statistician must acquire in order todo fruitful, work in statistics. Training andexperience in professional work at this levelcan only be given in a specialized institutionwhere such work is being actually done.

At a higher level, that is, at the stage ofbachelor’s or the master’s degree it should bepossible of course to cover a good deal of themore general part of the formal andmathematical theory together with a broadgeneral knowledge of statistical methods andapplications. General educational courses instatistics thus, in my opinion, can beconveniently given (in secondary schoolsin the first instance, and then) in theuniversities.

Professional training in statistics, however,requires a different type of institution. Theposition is somewhat analogous to medicalscience. Training in some of the basic subjectssuch as physics, chemistry, botany, zoology,physiology, etc. can be given in theuniversities. Professional training in medicalsubjects, however, requires specializedinstitutions working in close connection withclinical hospitals and surgeries. In the sameway professional training in statistics can begiven in an adequate manner only in closetouch with statistical projects. The skills andcraftsmanship, which I have described in thispaper, for example, can be more convenientlyacquired in statistical workshops rather thanin university classrooms.

But professional training is not confinedmerely to the acquiring of statisticalcraftsmanship but must go much further. Ihave no time to discuss this point at length. Inthe ultimate analysis professional competenceis determined by the ability

(a) to judge when the solution of a particularproblem or the making of anadministrative decision requires the helpof statistical information;

P.C. Mahalanobis with J.K. Galbraith, U.S. Ambassador to India, and

L.B. Johnson, then Vice President of the USA in New Delhi, May 20,

1961.

Statistics Day 2008

16

STATISTICS: A NEW TECHNOLOGY OF

MODERN AGE2

by Sir Ronald A. FisherFellow of Royal Society, London

It is now just about 25 years, since I firsthad the pleasure and privilege of meeting inIndia both Professor Mahalanobis and the veryable group of friends and students whom hehad gathered to form the Indian StatisticalInstitute. I had, indeed, before that knownsomething of Mahalanobis’ work, for I wasmuch concerned to view, if necessary onlyfrom a distance, the movements or thought indifferent parts of the world associated with therapid development of statistical under-standing in my own country; and I hadrecognised the appearance in the East of a newmovement which was putting India not farfrom the centre of the statistical map.

A movement comparable in importance,for example, with that of Yates in the extensionand elaboration of ExperimentalDesign, or of that which underthe impulse of Walter Shewhartwas soon to storm through theUnited States under the bannerof Quality Control. I needhardly say that I refer to theemergence of a statisticallycompetent technique of Sample Survey, withwhich I believe Professor Mahalanobis’ namewill always be associated.

What at first most strongly attracted myadmiration was that the Professor’s work wasnot imitative. That is a fault which has beenascribed to Indians, but as every statisticianknows nations are composed of very many,

and very differentindividuals. They donot share the same list offaults, and, at thepresent time you haveonly to look at anybunch of newly published books, on statisticsfor example, to see that this fault is by no meansa monopoly of any one people. Imitative booksare as common as dirt; and work of strikingoriginality is as rare now, as it was when I firstread of the surveys of the jute crop of Bengalin which the Professor was trying out his newideas.

During the inter-war period and indeedbefore, there had been some discussions at theInternational Statistical Institute on sampling

for the ascertainment ofdemographic and economic data.The conditions imagined werevery different from those ofpractical work in India, andindeed the theoretical principlesalso had been very imperfectlyappreciated. Even so elementary

a requirement as randomization was at firstignored. What was striking therefore in theIndian contribution to the problems of SampleSurvey, was that it combined a clear realizationof statistical principles with a down-to-earthexperience of the practical difficulties, in acountry in which education was, on the whole,backward, of getting such work done, on a

2Lecture delivered at the 1st Convocation of Indian Statistical Institute in February, 1962.

“..the accuracy of

sample surveys when

competently organised,

have now been put

beyond question”

“Imitative books

are as common

as dirt”

17

Statistics Day 2008

large scale, and yet with the primary scientificrequirement of ascertainable anddemonstrable precision. From that point thestandard set has never gone back, and publiclyorganised censuses and surveys in all parts ofthe world look to the Indian Sample Survey asa basis for comparison. Difficulties of coursethere are, in India, and also elsewhere, but thespeed and economy, as well as the accuracy ofsample surveys when competently organised,have now been put beyond question.

The needs of India,which are always, I fancyat the heart of theProfessor’s thinking, showthemselves again in theutilisation of thoseaptitudes in which Indianshave shown themselves tobe especially gifted. Fromtime to time as neweditions of Statistical Tables are required, I goover those contributions of sufficientimportance to be noticed in the preface,especially to combinatorial problems arisingfrom Yate’s work on experimental design andhis invention of balanced incomplete blocks.Mathematicians of many nations havecontributed to our knowledge of this veryintriguing field, but I donot think I exaggerate if Isay that the Indian namesare as numerous as allothers together, and thismust particularly beascribed to the fact thatearly in the Institute’shistory several of the young mathematiciansbrought into intellectual contact by theInstitute’s activity, found in this subject a typeof problem ideally suited to their gifts. Againthe important work in multi-factorial analysis

in which I have been glad to see extensionsand amplifications of my own early work, hasbeen chiefly, if not wholly taken up by pastand present members of the Institute such asBose and Roy in North Carolina, and especiallyby Rao here in Calcutta.

Every time I see the Institute it is biggerand I hope better. On this visit I am sure I havenot seen it all. It certainly shows no sign ofnarrowness or standardization. New facets are

always appearing-apaleontological museum ora garden of Indiancultivated plants. I takethese bold excursions tomean that the Directorbelieves, as I most certainlydo myself, that teaching,instruction or training instatistics, at whatever levelis bound to gravitate to an

exhibitionism in useless mathematics, unlessit is linked as intimately as may be, on the oneside with fact-finding projects in the traditionalstatistical fields of demography andeconomics, and on the other side withopportunities to gain first-hand familiaritywith at least some field in the natural sciences.Moreover, the science with which the student

is to become acquaintedmust be genuine research inits own right, not what iseloquently called a “mock-up” for the use of studentsonly.

Visitors are oftensurprised when they learn

that Miss Robinson and her colleagues areactually digging up some of the earlyinhabitants of India, and writing a new chapterin Indian mesozoic history. They ask what hasthat to do with statistics. That sort of question

“..teaching, instruction or training in

statistics is bound to gravitate to an

exhibitionism unless it is linked as

intimately as may be…with fact-

finding projects in the traditional

statistical fields of demography and

economics”

“Literary men and journalists“Literary men and journalists“Literary men and journalists“Literary men and journalists“Literary men and journalists

have often shown their aptitudehave often shown their aptitudehave often shown their aptitudehave often shown their aptitudehave often shown their aptitude

for getting hold of the wrongfor getting hold of the wrongfor getting hold of the wrongfor getting hold of the wrongfor getting hold of the wrong

end of the stick”end of the stick”end of the stick”end of the stick”end of the stick”

Statistics Day 2008

18

reveals well how comparatively narrow andtrivial a subject statistics was in the nineteenthcentury. It has grown since then, almostexplosively; and it is certainly proper, somewould say it is a prime requirement, if theStatistical Institute also widens its range ofinterests. The answer to the question, therefore,of what the sciences have to do with statisticslies in the part they must play in the educationof any competent statistician. And that theInstitute’s future stands or falls by the qualityof the education it offers.

The implementation of this broadeducational policy seems to me extremelydifficult. For its educational programme theInstitute needs not only leaders inmathematical thought like Professor Rao, whocan uphold and maintain the high place inworld opinion that Indians have already won,but they need also that patient and obstinateclass of scientific workers, who alone cantransmit the feel and know-how of the naturalsciences. There is very little of the aggressiveand self-advertising about such men: thoughthey may be among the most gifted of teachers,for whom even the best text books are no

substitute at all. I do not suggest that they areeasy to find, only that, here as elsewhere, theyare fitted to play the most important part inscientific, technological and technicaleducation.

These, may I insist, are not castes. Thoughliterary men and journalists have often showntheir aptitude for getting hold of the wrongend of the stick, by suggesting that technologistin particular are less than highly educated.Nothing could be further from the truth.Scientist, indeed may, at their own risk,narrow their interests in order to gain specialproficiency in one line. Technicians also maydevelop by ample practice very specialaptitudes. The technologist must talk thelanguage both of the scientist and of thetechnician. His education must be broader thantheirs, though at points less intensive.He has to see both sides of the fence, and is thechannel through which alone the skillsof the others can be made effective. It is, Ibelieve, in recognising Statistics as the keytechnology of our century, that we appreciatethe special features of the Indian StatisticalInstitute.

R.A. Fisher and Prasanta Chandraa signing the register on the occasion of

the award of medals in the city hall of Paris in September 1961

19

Statistics Day 2008

FISHER AND MAHALANOBIS: TWO PERSONALITIES

OF GREAT RESEMBLANCE

R. A. Fisher P. C. Mahalanobis

Born on Feb.17,1890 July 29, 1893

Died on July 28, 1962 June 28, 1972

(71 years) (79 years)

Father’s profession Business Business

Studies Cambridge, 1909-1913 Cambridge, 1913-1915

Math-Physics Physics

Influence and Mathematical Theory Biometrika

motivation of Evolution by journal edited by

Karl Peat1lon Karl Pearson

First publication 1912 1922

Editor Annals of Eugenics Sankhya

Active in statistics 50 years 50 years

Contributions *Statistics as a method *Statistics as a tool

in scientific research in increasing efficiency

(inductive inference) of all human efforts

*Experimentation for *Sample surveys for

comparative studies fact finding

Statistics Day 2008

20

STATISTICS: SCIENCE, TECHNOLOGY, OR ART3

As I have already said, statistics in theoriginal etymological sense is the science ofstatecraft, which represents the activity ofcollection and compilation of data and theiruse in policy making. The applications weremainly confined to economic data.

During the nineteenth century, statistics,which meant data, began to acquire a newmeaning as interpretation of data or drawingconclusions from data irrespective of theirsource. How do we conclude whether certainexperimental data support or contradict agiven theory? How can we make forecasts ofsocio-economic characteristics of a populationbased on current trends? Will it raintomorrow? What is the uncertainty in ourconclusions? Questions of this kind began tobe raised. But it is only in the beginning ofthe 20th century satisfactory answers could befound using the concepts of probability,actuarial science and theory of errors, whichbegan to be studied earlier more or lessindependently. Statistics developed as amethodology for extracting information fromdata and expressing the amount of uncertaintyin decisions we make.

Is statistics, as studied and practiced today,a science, technology or art? Perhaps it is acombination of all these.

It is a science in the sense that it has anidentity of its own with a large repertoire oftechniques derived from some basic principles.These techniques cannot be used in a routineway; the user must acquire the necessaryexpertise to choose the right technique in agiven situation and make modifications, if

necessary. Statistics plays a major role inestablishing empirical laws in soft sciences.Further, there are philosophical issuesconnected with the foundations of statistics –the way uncertainly can be quantified andexpressed - which can be discussedindependently of any subject matter. Thus ina broader sense statistics is a separatediscipline, perhaps a discipline of alldisciplines.

It is a technology in the sense that statisticalmethodology can be built into any operatingsystem to maintain a desired level and stabilityof performance, as in quality control programsin industrial production. Statistical methodscan also be used to control, reduce and makeallowance for uncertainty and therebymaximize the efficiency of individual andinstitutional efforts.

Statistics is also an art, because itsmethodology which depends on inductivereasoning is not fully codified or free fromcontroversies. Different statisticians may cometo different conclusions working with the samedata. There is usually more information ingiven data than what can be extracted byavailable statistical tools. Making figures telltheir own story depends on the skill andexperience of a statistician, which makesstatistics an art, as in the example of the RedFort Story4 .

What is the future of statistics? Statistics isnow evolving as a metascience. Its object isthe logic and methodology of the othersciences- the logic of decision making and thelogic of experimenting in them. The future of

3 Extract from “Statistics and Truth” by C. R. Rao, 19894 The story with the title ‘Salt in Statistics’ is also included in this Brochure

21

Statistics Day 2008

statistics is in communication of statisticianswith research workers in other branches oflearning; it will depend on the way theprincipal problems are formulated in otherfields of knowledge.

On the logical side, the methodology ofstatistics is likely to be broadened for usingexpert evidence in addition to informationsupplied by data in assessment of uncertainty.

Having said that statistics is science,technology as well as an art-the newlydiscovered logic for dealing with uncertaintyand making wise decisions-I must point out apossible danger to its future development. Ihave said earlier that statistical predictionscould be wrong, but there is much to be gainedby relying on statistically predicted valuesrather than depending on hunches orsuperstitious beliefs. Can the customer forwhom you are making the prediction sue youif wrong? There have been some recent courtcases. I quote from the editorials of ThePittsburgh Press.

Saturday, May, 1986Forecasters Breathe Easier

A federal appeals court has wisely corrected agross miscalculation of government liability in acase involving weather forecasting.

Last August, a U.S. District judge awarded $1.25 million to the families of three lobster-men whowere drowned during a storm that had not beenpredicted. The judge said the government wasliable because it had failed to repair promptly a windsensor on a bouy used to help forecast weatherconditions off Cape Cod.

The award was overturned the other day by theappeals court on grounds that weather forecastingis a “discretionary” function of government andnot a reliable one at that.

“Weather predictions fail on frequentoccasions” the appeals court said. “If in onlyproportion parties suffering in consequencesucceeded in producing as expert who couldpersuade a judge…….that the government shouldhave done better, “The burden on the government”would be both unlimited and intolerable”.

The case isn’t over yet, since it probably will beappealed to the Supreme Court. But governmentmeteorologists practicing their inexact science arebreathing a bit easier.

“...Statisticians working in the Government and in Industry are often faced with languagebarriers with their bosses. The chief of a statistical office, an officer in administrative service,was meeting a group of statisticians who complained that in a report received from anotherorganization some estimates were given without the standard errors. The chief was reportedto have immediately remarked, ‘Are there standards for errors too?’

A report submitted to a Tea Board by a consulting statistician contained a table withthe caption: Estimated number of people taking tea with standard error, economic status.Soon a letter was sent to the statistician asking what kind of snack standard error was,which people take with tea.”

Statistics Day 2008

22

SALT IN STATISTICS5

“……and, what is still more extraordinary, I have met with a philosophical work in which the utilityof salt has been made the theme of an eloquent discourse, and many other like things have had a similarhonour bestowed on them.”

Pheadrus (Plato’s Symposium on Love)There were communal riots in Delhi in 1947immediately after India achievedindependence. Large number of people of aminority commu8nity took refuge in the RedFort which is a protected area, and a smallnumber in the Humayun Tomb, another areaenclosing an ancient monument. TheGovernment had the responsibility to feedthese refugees. This task was entrusted tocontractors, and in the absence of anyknowledge about the number of refugees, thegovernment was forced to accept and pay theamounts quoted by the contractors for differentcommodities purchased by them to feed therefugees. The government expenditure on thisaccount seemed to be extremely high and itwas suggested that statisticians (who count)may be asked to find the number of refugeesinside the Red Fort.

The problem appeared to be difficult underthe troubled conditions prevailing at that time.A further complication arose as the statisticalexperts called in to do the job belonged to themajority community ( different from that of therefugees) and their safety could not beguaranteed if the statistical techniques to beapplied by them for estimating the number ofrefugees required their getting inside the RedFort. Then the problem before the experts wasto estimate the number of persons inside agiven area without any prior information aboutthe order of magnitude of the number, without

having any opportunity to look at theconcentrations of persons inside the area andwithout using any known sampling techni-ques for estimation or census methods.

The experts had to think of some way ofsolving the problem. Giving up might beinterpreted by the government as failure ofstatistics and/or of the statisticians. They had,however, access to the bills submitted by thecontractors to the government, which gave thequantities of various commodities such as rice,pulses and salt purchased by them to feed therefugees. They argued as follows.

Let R, P and S represent the quantities ofrice, pulses and salt used per day to feed allthe refugees. From consumption surveys, theper capita requirements of these commoditiesare known, say, r, p and s respectively. ThenR/r, P/p and S/s must provide parallel(equally valid) estimates of the same numberof persons. When these ratios were computedusing the values R, P and S quoted by thecontractors it was found that S/s had thesmallest value and R/r the largest valueindicating that the quantity of rice, which isthe most expensive commodity compared tosalt, was probably exaggerated. (The price ofsalt was extremely low in India at that timeand it would not pay to exaggerate the amountof salt). The estimate S/s was proposed bythe statisticians for the number of refugees inthe Red Fort. The proposed method wasverified to provide a good approximation to

5 Extract from “Statistics and Truth” by C. R. Rao, 1989

23

Statistics Day 2008

the number of refugees in the Humayun tomb(the smaller of the two camps with only arelatively small number of refugees), whichwas independently ascertained.

The salt method arose out of an ideasuggested by the late J.M. Sengupta who wasassociated with the Indian Statistical Institutefor a long time. The estimate provided by thestatisticians was useful to the government in

taking administrative decisions. It alsoenhanced the prestige of statistics whichreceived good government support ever since,for the development in India.

The method used is unconventional andingenious, not to be found in any text book.The idea behind it is statistical reasoning orquantitative thinking. Perhaps, it also involvesan element of art.

United Nations Statistical Commission, New York, 16 April-3 May, 1956, with P.C. Mahalanobis

in the chair

Statistics Day 2008

24

Blank

Statistics Day 2008

26

Blank

27

Statistics Day 2008

PM’S INAUGURAL ADDRESS AT ‘STATISTICS DAY’

CELEBRATION JUNE 29, 2007 NEW DELHI

“I am delighted that we are commemo-rating the birth anniversary of Late ProfessorPrasanta Chandra Mahalanobis as StatisticsDay in India. This is a tribute to his enduring,pioneering and multifaceted contributions tothe discipline of statistics. He was a worldleader in the discipline. Along with ProfessorR.A. Fisher, Professor Mahalanobis is rightlycredited for giving a distinct status to statisticsas an independent scientific discipline andplacing it among the world of modern sciences.

Professor Mahalanobis was not merely agreat academician. He never accepted statisticsas a purely academic discipline meant only forformal teaching and research. He conceived itas an essential and integral tool for guidingsocial and economic development. To put it inhis own words:

“If the problem is one of theoretical nature,statistics supplies a valid method for drawinggeneral conclusions from particularexperience. If the problem is a practical one,statistics supplies the basis for choosing aparticular course of action (in preference toother possible courses) by balancing the risksof gain and loss.”

Prof. Mahalanobis believed that it wasimportant to develop innovative statisticaltechniques appropriate to each social andeconomic sphere. In doing so he emphasizedthe universal applicability of tools of statisticsin all major scientific disciplines. Hispioneering contributions encompassed a widerange of areas of practical application. Theseincluded anthropology, educational testingand statistical quality control, on the one hand,and agriculture, flood control and meteorologyon the other. There is indeed, practically no

aspect of human endeavour in this country thathis genius did not touched, directly or throughthe people he inspired.

Professor Mahalanobis was also a greatinstitution builder. By founding the IndianStatistical Institute in 1931, and starting theworld famous journal ‘Sankhya’ in 1933, heprovided an institutional umbrella forcatalyzing the spread of teaching and researchin statistics. India today can proudly boast ofhaving a distinct Indian School of Statistics –an achievement that few other scientificdisciplines enjoy in our country. Certainlyindividual scientists have made path-breakingcontributions in other disciplines, but in nonehave they created a special niche for Indiawithin the international order, as ISI has donein the field of statistics.

Professor Mahalanobis’s recognition of theimportance of institutional structures and hisall-encompassing vision of statistics were alsoresponsible for the active interest he took indeveloping the official statistical system in ourcountry. He was appointed HonoraryStatistical Adviser to the Union Cabinet in 1949and was the Chairman of the first NationalIncome Committee in 1950. He conceptualizedand brought into existence, the two majorpillars of the official statistical system in ourcountry today, namely, the Central StatisticalOrganisation (CSO) for systematization andcollection of administrative data and theNational Sample Survey Organisation (NSSO)for conducting large scale sample surveys as aguide to policy planning. Indeed, the NSSOhad its origin in the Indian Statistical Instituteand became a part of the official statisticalsystem only much later.

Statistics Day 2008

28

Prof. Mahalanobis also recognised that thedecentralized statistical system in India and thefederal principles of our polity require effectivesystems of coordination for the collection ofstatistics in such a large country. He, therefore,organized the Inter-Ministerial StandingCommittee in the Cabinet Secretariat and thebiennial Conference of Central and StateStatistical Organisations, which providedsound institutional mechanisms for horizontaland vertical coordination of statistical activitiesin our country. At the international level, hecontributed to the establishment of the UnitedNations Statistical Commission.

We celebrate Statistics Day today not onlyto honour the memory of this great man, butalso to give due recognition to the fact thatstatistics affects and influences all majordecisions we take, big or small in public life.Often in ways we may not even realize. As weprogress in our path of development towardsbecoming a knowledge economy, and as weintegrate with the international economicorder, the role of statistics will becomeprogressively more important. In such asituation, the hallmark of a country’sreputation will be the credibility andtransparency of its statistical system and thestatistical data emulating from that system.

We are indeed fortunate that over the yearswe have been able to maintain the highstandards of integrity in our official statistics.However, in such matters perceptions are asimportant as the reality. It is for this reasonthat our government took a major step last yearby bringing into existence the NationalStatistical Commission, which is anindependent body of professionals, for guidingthe official statistical system. I am very happythat my esteemed friend, Professor Tendulkaris the Chairman of this highly prestigiousnational institution.

Its arms length relationship with the

Government should give our official statisticsenhanced international credibility andtransparency. It should also help improve thelevel of confidence that people andorganizations have in the confidentiality of thedata that they provide, thereby improving boththe completeness and the accuracy of the dataso collected.

Of the wide-ranging mandate that we havegiven to the National Statistical Commission,two are of particular importance. The first isto evolve and lay down national qualitystandards in different fields of statistics; andthe second is to exercise statistical coordinationbetween central Ministries, Departments aswell as State Governments. I appeal to allconcerned to extend fullest possiblecooperation to the National StatisticalCommission and to actively seek its assistancein improving our statistical data bases.

Official statistics, of course, is only one partof the wide variety of data that is required bya modern dynamic economy. This StatisticsDay is also meant to recognize the contributionof many statisticians who work in variousacademic institutions, private enterprises, nonprofit organizations and other institutionsdealing with quality control.

They all provide essential informationrequired for optimal decision making indiverse fields. We should all be proud of ourcollective heritage in the field of statistics andresolve to take our country to the pinnacle ofthis very important discipline. I am certain thatour young statisticians will continue to leadthe world in theoretical and applied statistics,and bring a further sense of pride andaccomplishment to our Nation. That will be abefitting tribute to the memory of Prof. P.C.Mahalonobis. I therefore feel a great sense ofpride to honour Professor Mahalonobis bynaming his birthday as ‘Statistics Day’.

Thank you.”

29

Statistics Day 2008

STATISTICAL SYSTEM OF INDIA: SOME REFLECTIONS

Rakesh Mohan*

I am delighted and honoured to be at thisauspicious occasion of the first StatisticsDay and to inaugurate first Annual Conferenceon Financial Statistics. As a member of thestatistics community, it is matter of greatpride that June 29, birth anniversary of (late)Professor Prasanta Chandra Mahalanobis, hasbeen declared by the Government of India asStatistics Day. I am particularly happy thatProfessor Parikh is here who began hisprofessional life in the ISI Statistical Unit in thePlanning Commission and had a great deal ofinteraction with Professor Mahalanobis.

Statistics in simple terms is defined as thestudy of the ‘laws of chance’. On this StatisticsDay, let me however not take a chance to talkabout field statistics. In the preface to his bookon “Statistics and Truth”, C.R.Rao (1989)writes, which I thought is worth quoting atlength,

“Statistics as a method of learning fromexperience and decision making underuncertainty must have been practiced from thebeginning of mankind. But the inductivereasoning involved in these processes havenever been codified due to the uncertain natureof the conclusions drawn from given data onformation. The breakthrough occurred only inthe beginning of the present century with therealization that inductive reasoning can bemade precise by specifying the amount ofuncertainty involved in the conclusions drawn.This paved the way for working out an optimumcourse of action involving minimum risk, in

any given uncertain situation, by a purelydeductive process. Once this mechanism wasmade available, the floodgates opened and therewas no end to the number of applicationsimpatiently awaiting for methods which couldreally deliver the goods. From the time ofAristotle to the middle of the 19th century,chance was considered by scientists as well asphilosophers to be an indication of ourignorance which makes predictions impossible.It is now regarded that chance is inherent in allnatural phenomena, and the only way ofunderstanding nature and making predictions(with minimum loss) is to study the laws (orthe inner structure) of chance and formulateappropriate rules of action. Chance may appearas an obstructor and an irritant in our dailylife but chance can also create. We have nowlearnt to put chance to work for the benefit ofmankind. All knowledge is, in the final analysis,history. All sciences are, in the abstract,mathematics and all methods of acquiringknowledge are essentially statistics.”

What has been the historical developmentof the statistical system in India? Who havebeen its main architects? What has been therole of the Reserve Bank in the nationalstatistical system? I thought these questionswill be of interest to the audience assembledhere today.

Historical developments

Let me go back in history. India has a longhistorical tradition of collection and use of

*Inaugural address by Deputy Governor, Reserve Bank of India on the Statistics Day and Annual Conference on FinancialStatistics, on June 29, 2007 at Department of Statistical Analysis and Computer Services, Reserve Bank of India , Mumbai.Assistance of A K Ray and Abhiman Das in preparing the speech is gratefully acknowledged.

Statistics Day 2008

30

various kinds of statistics. Kautilya’sArthasastra (321-296B.C.), one of the greatesttreatise of economics, indicates a system ofcensus and data collection relating toagriculture, population and other economicactivities, covering villages and towns. Inaddition, the concept of crosschecking andvalidation by independent agents was verymuch part of the data collection system. Thus,early evidence established the coexistence ofeconomics and statistics, and their applicationfor the welfare of the states. Subsequenthistorical anecdotes based on Huen Tsang’swritings (dated late seventh to early eighthcentury) give a detailed description of the planof cities, construction of houses, and an accountof common products of India and data on thearea of kingdoms and the distances betweenthem. During the Moghul period, evidence andapplication of statistical knowledge wasprominent in Ain-i-Akbari by Abul Fazal.Documentary evidence includes the system oflegalised measurements, land classificationand crop yields by season, etc. The system ofland tenure and land revenue, followed duringthe Moghul period, had enough empirical basis.

The statistical system was strengthenedduring the British period. During this period,the statistical development was geared

towards administration, tax collection,revenue, trade and commerce and relatedactivities as might be expected. With a view torationalize and establish a sound system ofrevenue collection, the East India Company in1807 emphasized the need of a statisticalsurvey of the country. A small department ofstatistics was founded in the India House in1847. In 1848, the first census relating to thearea and revenue of each in North-WestProvinces was released. In 1853, thedepartment released the first series of statisticalpapers on India. Impressed by the trend instatistical activities, the Secretary of Stateordered the Governor- General in Council toprepare a ‘comprehensive and coordinatedscheme of statistical survey’ for each of thetwelve great provinces of the then British Indiaand Dr. W.W. Hunter was appointed asDirector-General of Statistics in India in 1869,who can perhaps be regarded as the originalprecursor to the Chief Statistician of Indiatoday. The Statistical Account of Bengal (thepresent Bangladesh, West Bengal, Bihar andOrissa) was published in 20 volumes. For eachdistrict there were details on topographicaldata, ethnic divisions and creeds, agriculturalsituation, commerce, working of districtadministration and finally the sanitary andhealth aspects. Although censuses of Calcuttawere undertaken in the year 1822 and in 1847,the operation of a decennial census for thewhole country started in 1881 and iscontinuing ever since. The report on theCensus of British India taken in 1881 waspublished in three volumes.

The need for timely and accurate collectionof agricultural data was felt by the IndianFamine Commission and agriculturaldepartments were organised in variousprovinces which resulted in the publicationof ‘Agricultural Statistics of British India’ in

Rajendra Prasad, President of India, speaking at the inauguration

ceremony of the International Statistical Conference, Rastrapati

Bhavan, New Delhi, 5 December, 1951, in presence of P.C. Mahalaobis

31

Statistics Day 2008

1886. To scrutinise and summarise the datacollected by the agricultural departments, astatistical Bureau was formed at the centre in1895 to coordinate the agricultural, foreigntrade, prices, wages and industrial statistics.During 1905, a separate body DirectorateGeneral of Commercial Intelligence andStatistics (DGCI & S) was constituted tocollect/publish commercial and trade statisticsand to help trade and business. The first issueof Indian Trade Journal was released in 1906and first price statistics based on a survey wasreleased in 1910. The Economic EnquiryCommittee set up in 1925 under theChairmanship of Dr. Visweswarayya andmore importantly the Bowley- RobertsonCommittee set up later in 1934, were mainlyresponsible for the government’s decision toset up an Inter-Departmental Committee withthe Economic Adviser to the Government ofIndia as the chairman. The Inter-DepartmentalCommittee recommended the formation of aCentral Statistical Office for coordination,institution of a statistical cadre, establishmentof State Bureaus at State Head Quarters andmaintenance of important statistics for theentire country.

Architect of Modern Statistical Methods inthe Indian Subcontinent

The developments in statistics that tookplace between 1930 and 1960 are quiteremarkable and in some sense unique. Noother discipline in India recorded such growthand development during the same period inIndia. There were several importantingredients. In statistics, unlike otherdisciplines, India was not a late starter. Indeed,much of the development even in the UnitedStates came later. Only Britain had startedearlier. This helped creation of an Indian schoolof statistics with its own mix of theory andapplications. The architect of modern statistical

methods in the Indian subcontinent wasundoubtedly Professor Prasanta ChandraMahalanobis. He was helped by a verydistinguished group of scientists that includedR.C. Bose, S.N. Roy C.R. Rao, S.S. Bose, K.R.Nair, D.B. Lahiri and many others. There wereothers like P.V. Sukhatme, and V.G. Panse whoworked independently of Mahalanobis. Thehistory of statistics is basically a history ofsome of these persons as well as a history ofinstitutions and interactions between personsand institutions. The institute around whichall these development took place is none otherthan the Indian Statistical Institute (ISI),Calcutta. Looking back at the history, oneobserved that 1950s was inarguably the goldenperiod of ISI. Besides, Prof. Mahalanobis andC. R. Rao, the faculty of ISI included R.R.Bahadur, D. Basu, G. Kallianpur, D.B. Lahiri,M. Mukherjee, R. Mukherjee and many otherdistinguished luminaries. Some of themstudied abroad and subsequently joined ISI asa faculty. D. Basu joined the ISI as a studentand later became a Professor as well as the firstDean of Studies. D. B. Lahiri was a self-taughtmathematician, contributed significantly onsample survey methods. Between them Rao,Bahadur, Basu and Kallianpur and a newgroup of brilliant students including K.R.Parthasarathy, R. Ranga Rao, V.S. Varadarajan,S.R.S. Varadhan made fundamental

Professor P.C. Mahalanobis in a field visit in Giridih Branch, in 1953

Statistics Day 2008

32

contributions to probability and classicalinference during this period which were asimportant as the earlier contributions of Boseand Roy to design of experiments andmultivariate analysis. Among the other earlierstudents of the Institute who achievedinternational reputation are G.P. Patil,T.N.Srinivasan, R.G. Laha, J. Roy, Sujit KumarMitra, D.K. Roy Choudhury, and others. I mayalso mention that S.R.S. Varadhan, currently aprofessor of mathematics at New YorkUniversity and only Indian ever, won theprestigious Abel prize (which is perhapsequivalent to a Nobel Prize) in mathematicsthis year.

Professor P. C. Mahalanobis and Statistics

As the Statistics Day is being celebrated inhonour of the birth anniversary of (late)Professor Prasanta Chandra Mahalanobis, Ithought a brief sketch of his life will beappropriate at this juncture. Born on June 291893 in a well-to-do progressive Brahmofamily, he passed away on June 28, 1972. Hegraduated in physics from Presidency College,Kolkata and went on to study in Cambridgein 1913 and in 1915 finished his Tripos in

Natural Science with a first class. His firstencounter with Statistics was also a chance-event, when his tutor Macaulay drew hisattention to some bound volumes of Biometrika.He got so interested that he bought a completeset of Biometrika volumes towards his journeyback to India. Mahalanobis set up the StatisticalLaboratory in the Presidency Collegesometime in the 1920s. His contribution tostatistics is enormous. During the 1920s anduntil the mid-1930s, all or nearly all thestatistical work done in India, was done single-handedly by Mahalanobis. The early statisticalstudies included analyses of data on statureof Anglo-Indians, meteorological data, rainfalldata, data on soil conditions, etc. Some of thefindings of these early studies were of greatimpact in the control of floods, developmentof agriculture, etc., and led to the recognitionof Statistics as a key discipline. His passion andconviction was truly guided for the solutionsof applied problems. His further work onanthropological data resulted in newmethodology for classifying or distinguishingpopulations characterized by suchmeasurements called Mahalonobis D-square.1

Another very important initiative undertakenby Professor Mahalonobis was the initiationof Crop Cutting Surveys to estimateagricultural production in India, which was anovel idea of that time. These surveys havecontinued till today and continue to beimportant in the estimation of agriculturalproduction. The Indian Statistical Institute wasfounded in a small room in the PresidencyCollege as a society on 17 December, 1931.Sankhya, the Indian Journal of Statistics, wasfounded two years later. Mahalanobis’influence was so ubiquitous that even studentsof Physics began to take interest in statistics.Later, several talented young scholars joined

1 He submitted the paper in Biometrika for publication. Pearson had expressed reservations about Mahalonobis D-square andthus did not publish it. However, Mahalanobis published it in other place.

P.C. Mahalanobis with Professor Haris and Meghnad Saha

at Amrapali, in ISI, in 1953

33

Statistics Day 2008

to form an active group of statisticians.Mahalanobis continued to be the nucleus.Theoretical research in Statistics began toflourish in all the areas of statistics includingdesign of experiments, statistical qualitycontrol, operations research, etc. Research onlarge scale sample surveys won Mahalanobisa Fellowship of the Royal Society. Design andanalysis of agricultural experiments alsobloomed and led to some internationalcontacts, notably with Sir Ronald A. Fisher. Atthe request of the Government of India, a draft

award degrees and diplomas in Statistics. Thefirst computer of the country was installed inISI in 1960. Subsequently, ISI made significantcontributions to the development of advancednumerical algorithms.2 Mahalanobis’ masteryover empirics was exemplary as illustrated inusing the method of fractal graphical analysisof consumption data, which was published inEconometrica (1960).3 It is also understood thatthe concept of interpenetrative sub-samplingand pilot survey, pioneered by Mahalanobis,essentially form the basis of sequential analysisand modern bootstrap methods.

Evolution of the Statistical System afterIndependence

After independence in 1947, the countrysaw an urgent need for a statistical frameworksuitable for economic and social development.Mahalanobis was appointed as an HonoraryStatistical Adviser in 1949 to the Governmentof India and Central Statistical Unit was setupunder his technical guidance which was laternamed as Central Statistical Organization(CSO) in 1951.This organization was setupmainly to coordinate the statistical work donein various ministries and other governmentagencies and to advise them, to maintainstandards with regard to definitions, conceptsand procedures, to provide consultancy, toliaison with international statisticalorganizations, to prepare and publish aMonthly Statistical abstract and an AnnualStatistical Abstract and to inform annualstatistical information to public. In India, priorto 1947, the estimation of national income wasattempted by individual economists andscholars for specific years. Among these, themost systematic work was that of V.K.R.V. Raoin his book National Income in British India 1931-

2 It is also believed that the seed of computer and IT revolution in India was germinated in ISI.3 The editor of Econometrica published a shorter version of this paper what reportedly Mahalanobis did not like. As a result, he

again published the full version of the paper in Sankhya.

Pandit Jawaharlal Nehru, Prime Minister of India, inaugurating stud-

ies on national planning at I.S.I., Calcutta, on 3 November 1954

of the second Five Year Plan was prepared atthe Indian Statistical Institute in 1954 byMahalanobis. The second plan pattern ofindustrial investment, with its marked shift infavour of capital good industries, was deeplyinfluenced by the two-sector growth modeldeveloped by Mahalanobis. The substantialcontributions of the Institute to theoretical andapplied work, its training and promotionalactivities culminated in recognition by theGovernment of India. The Parliament passedthe Indian Statistical Institute Act, 1959 whichdeclared the Institute as an “Institution ofNational Importance” and empowered it to

Statistics Day 2008

34

32 (London; MacMillan 1940), which formedthe basis of national income estimation in thepost-independence period. In 1949, theGovernment of India formed the NationalIncome Committee (NIC) under theChairmanship of P.C. Mahalanobis, withV.K.R.V. Rao and D.R. Gadgil as members.From then onwards the national incomeestimation has been steadily strengthened.NIC recommended the holding of an annualconference on national income and wealth. TheFirst Indian Conference on Research inNational Income was organised by the CentralStatistical Organisation (CSO) in early 1957. In1964, this conference was converted into theIndian Association for Research in NationalIncome and Wealth (IARNIW). Dr. V.K.R.VRao was the first President of IARNIW. He iscredited with the building of IARNIW as anindependent research body under the aegis ofthe Central Statistical Organization. In orderto improve the quality and fill up the gaps instatistical information, it was decided toestablish a National Sample SurveyOrganisation (NSSO). This organization hadfour divisions: Survey design and research,Field operations, Data processing, andEconomic Analysis. The national samplesurvey (NSS) is the largest multi-purposesocio-economic survey.

The coverage and scope of data collectionand dissemination of CSO has increasedmanifold in recent times. Its current activitiesinclude National Income Accounting, conductof Annual Survey of Industries, EconomicCensuses and its follow up surveys,compilation of Index of Industrial Production,as well as Consumer Price Indices for UrbanNon-Manual Employees, HumanDevelopment Statistics, Gender Statistics,imparting training in Official Statistics, FiveYear Plan work relating to Development ofStatistics in the States and Union Territories;

dissemination of statistical information, workrelating to trade, energy, construction, andenvironment statistics, revision of NationalIndustrial Classification, etc.

The National Statistical Commission (NSC)set up by the Government in January 2000under the Chairmanship of Dr. C. Rangarajanreviewed the statistical system and the entiregamut of Official Statistics in the country. Corefunctions of NSC, inter alia, includeidentification the core statistics which are ofnational importance and are critical to thedevelopment of the economy, to evolvenational policies and priorities relating to thestatistical system and to evolve standardstatistical concepts, definitions, classificationsand methodologies in different areas instatistics and lay down national qualitystandards on core statistics. One of the keyrecommendations of this Commission was toestablish a permanent National Commissionon Statistics to serve as a nodal andempowered body for all core statisticalactivities of the country, evolve, monitor andenforce statistical priorities and standards andto ensure statistical coordination among thedifferent agencies involved. In line with theabove recommendations, the Government ofIndia ordered the setting up of a permanent

P.C. Mahalanobis with the statisticians of C.S.O. during a reception at

his New Delhi residence, 8 King George Avenue, 3 December, 1954

35

Statistics Day 2008

National Statistical Commission (NSC) on 1stJune 2005. The present setup of NSC is a typeof umbrella organization covering both CSOand NSSO.

National Statistical System and the Role ofthe Reserve Bank

Reserve Bank of India (RBI) collects andanalyses statistics on various economictransactions of banking and other financialinstitutions in the process of implementing itspolicies towards achieving its implicit twinobjectives of growth and price stability. Amajor part of the statistics in RBI is collectedthrough either statutory or control returns,which are exclusively used for monetary policyand supervision. In the field of informationgeneration and analysis we adhere tointernational standards and practices. Tocollect supplementary statistics, RBI conductssurveys in the fields of credit to rural/urbanhouseholds, industrial outlook, inflationexpectations, different aspects of bankingsector, external sector and the privatecorporate sector.

At present, the surveys conducted by RBI canbe broadly classified into five categories:

(a) external sector including (1) survey offoreign liabilities and assets for corporate,

insurance & mutual fund sectors, (2)coordinated portfolio investment survey,(3) survey on software export, (4)unclassified receipt survey used for BoP,(5) survey on balances in Nostro / Vostroaccount used in BoP, and (6) survey onnon-resident deposits;

(b) banking sector including (1) survey ondistribution of credit, deposits andemployment in banks, (2) survey oncomposition and ownership of depositswith scheduled commercial banks, (3)survey on investment portfolio ofscheduled commercial banks, (4) surveyof debits to deposit accounts withscheduled commercial banks, (5) surveyon international assets and liabilities ofbanks and (6) survey of small borrowalaccounts;

(c) corporate sector including survey ofperformance of private corporate businesssector, conducted since 1951-52;

(d) monetary policy including (1) industrialoutlook Survey, (2) inflation expectationsurvey for households and (3) survey ofinventories, order books and capacityutilization;

(e) ad hoc: Census of non-banking financialcompanies not accepting public deposits.

In addition to the abovementioned surveys,recently RBI has initiated steps to make use ofsurvey methods for various operationalpurposes, e.g., to assess the extent of successand realities of implementation of financialinclusion. You may be aware that in India,Government as well as the central bank attachenormous importance towards financialempowerment of the rural poor and RBI hastaken steps for financial inclusion and spreadof Self Help Groups. In support of monetary

United Nations Statistical Commission, New York, 16 April-3

May, 1956, with P.C. Mahalanobis in the chair.

Statistics Day 2008

36

policy, we propose to introduce survey ofprofessional forecasters, as done in most of thecentral banks in developed countries. Inaddition, it may be necessary to collectnecessary information for compiling housingstarts, in view of the increasing importance ofhousing construction in the economy.However, in view of the current state of dataavailability with local authorities, this willneed considerable additional work before anysignificant progress can be made. OtherCentral Banks conduct a number of othersurveys, both ad hoc and regular ones, in orderto help more informed monetary policymeeting, and other needs as they arise. As weimprove the technical basis of our monetarypolicy apparatus we will need to keepinnovating in this area.

Traditionally RBI is directly involved innational accounts compilation, especially insavings estimation and flow of fundcompilation. Basic information on householdand corporate savings is provided to CSO forpreparing the final estimates. Besides, data onoutput of banking system in GDP is imputedfrom the data supplied by us. RBI also has along tradition of conducting important surveysin collaboration of NSSO. The firstcomprehensive survey conducted by RBI isknown as All-India Rural Credit Survey, with1951-52 as the reference period. The objectiveof the survey was to collect such data/information as would assist the RBI and theGovernment of India in formulating anintegrated credit policy for rural credit and toassess the extent of indebtedness of ruralhouseholds to financial institutions in theorganized and unorganized sectors. Findingsof this landmark survey subsequently changedthe entire landscape of Indian banking andculminated in introducing the world’s biggestsocial banking experiments in the form of rural

credit and priority sector lending. Such All-India surveys were conducted decennially andthe latest survey pertained to the year 2002-03. While the 1951-52 and 1961-62 surveyscovered only rural households, the subsequentsurveys covered urban households also. Thesurveys of 1971-72 and 1981-82 wereconducted jointly with NSSO, Government ofIndia, whereas those of 1991-92 onwards wereconducted entirely by NSSO.

The other area, which has directconsequence with RBI activities, relates tomeasurement of output and price. As memberof the Technical Advisory Committee ofNational Accounts Statistics (NAS), Statisticson Prices and Cost of Living (SPCL), revisionof WPI and CPI series, etc., RBI has beenplaying a critical role. Let me highlight a fewrecent initiatives. The procedure of convertingcurrent price estimate into constant pricefollowed earlier, especially for estimating theoutput of financial sector, resulted in somekind of anomalies. In some cases, whileestimate at current price was negative,constant price estimate was positive andincreasing. However, in the last GDP revision(1999-00 series), this problem was corrected

Smt. Indira Gandhi, Prime Minister of India, delivering her

speech on the occasion of Golden Jubilee Celebration of the

Institute in Calcutta. Prof. C.R. Rao and Prof. Kallianpur are

on the dais along with others, 20th January 1982.

37

Statistics Day 2008

after a series academic discussion with CSO.Similarly, treatment of some of the financialintermediaries like mutual funds was notdefined clearly in SNA 1993. And treatingmutual funds similar to banks wasconceptually not correct. Recently, CSO set upa committee in RBI to look into this matter andthe recommendations of the committee hasbeen accepted by TAC of NAS. Besides, RBI isalso working in close collaboration with CSOfor implementing the ensuing changes of SNArevision due this year.

Let me cite an example of RBI’s recentinvolvement in price measurement issues. The

need for integration CPIs is nothing new. TheNational Statistical Commission (2001)recommended, inter alia, that the current CPIsdo not provide changes in the prices for theentire rural and urban population since theyare designed to measure the changes in theprices of goods and services consumed byspecific segments of the population and hencethere is a need to compile the CPI separatelyfor the entire rural and urban population. Inthis context, a Sub-Group set-up in RBI,recommended the compilation of separateCPIs for whole rural and urban population andprovided a roadmap for compilation. I happyto note that TAC on SPCL has finally acceptedthis and shortly we shall have CPI(U) andCPI(R). Needless to mention that these priceindices will be very useful for monetary policypurposes.

Application of Statistical Methods in RBI:

Need for Further Work

Before I conclude, let me outline a few areasof work in RBI where statisticians cancontribute significantly. First, as monetarypolicy formulation largely depends on theforwardlooking behaviour of the economy, theuse of forecasting techniques automaticallycomes in the forefront. While traditional timeseries techniques are useful, we need to buildup expertise on forecasting using calibratedmodels. Such models have definitiveadvantage of incorporating economic agentsutility, requires less data and can bemaneuvered easily with different alternatives.Second, RBI incurs a huge expenditure onmanagement of currency through out thecountry. Operations research techniques couldbe utilized effectively in the areas of inventory,transportation, etc., so as to develop an efficientand cost effective currency management

J.B.S. Haldane on his birthday with P.C. Mahalanobis at Amrapali,

5th November, 1957

Statistics Day 2008

38

systems. Third, risk analysis under a statisticalframework is a pre-requisite for implementingBasel II norms. Fourth, innovative samplesurveys in the area of financial inclusion,development of leading indicators likehouse-start index, inflation expectation, etc.,could be very useful tools. Fifth, we do nothave an appropriate measure of potentialoutput. It is not an easy task; however, effortsshould be made to devise a suitablemethodology for estimating potential outputat the earliest.

Finally, let me conclude by saying thatknowledge building is an ongoing activity.Frontiers are shifting upwards consistentlyand thus demanding a persistent catching upwith the advanced knowledge. Centralbanking today is more challenging than everand accurate information is the key tosupplement policy decisions. I hope theStatistics Day and the conference on financialstatistics pave the platform of knowledgesharing and understanding of advancedfrontiers.

39

Statistics Day 2008

Statistics Day 2008

40

Blank

41

Statistics Day 2008

AWARDS AND THE RECIEPIENTS

Prof. P.V. Sukhatme National Award in Statistics for Senior Statisticians

S. No. Year Name of the Awardees/Recipients

1 1999-2000 Prof. J.K. Ghosh, former Director, ISI, Kolkata.

2 2001-2002 Prof. Prem Narain, former Director, IASRI, New Delhi.

3 2003-2004 Prof. Belvant Keshav Kale, former Professor and Head of theDepartment of Statistics, University of Pune.

4 2005-2006 Jointly (i) Dr. Manindra Nath Das, former Professor and Director,IARS, New Delhi (ii) Prof. Nikhilesh Bhattachrya, formerProfessor, ISI, Kolkata.

5 2007-2008 Prof. B.L.S. Prakasa Rao, Jawaharlal Nehru Chair Professor,University of Hyderabad.

Prof. C.R. Rao National Award in Statistics for Young Statisticians

S. No. Year Name of the Awardees/Recipients

1 2000-2001 Jointly (i) Dr. Rahul Mukherjee, Professor, IIM,Kolkata (ii) Dr. Rajeeva Laxman Karandikar,Professor, ISI, New Delhi

2 2002-2003 Professor Arup Bose, ISI, Kolkata.

3 2004-2005 Prof. Probal Choudhuri, ISI, Kolkata

4 2006-2007 Prof. Ayanendranath Basu, ISI, Kolkata

International Award in Memory of Professor P. C. Mahalanobis

S. No. Year Name of the Awardees/Recipients

1 2003 Prof. C. R. Rao, Former Director, Indian Statistical Institute, India.

2 2005 Prof. Ben Kiregyera, Professor, Makerere University, Uganda

3 2007 Prof. I.P. David, Former Manager, Statistics and Data SystemDivision, ADB, Manila, Philippines.

Statistics Day 2008

42

Awards for “Essay Writing” on the occasion of Birth Anniversary ofProf. P. C. Mahalanobis

S. No. Year Name of the Awardees/Recipients Prize

1 2005 Sh. Mathew Francis, St. Thomas College, FirstDepartment of Statistics, Kottayam

2 2005 Ms. Sumi A. R., University of Kerala, Trivandrum Second

3 2005 Ms. Prerana Bhasker Ail, Department of Post and ThirdGraduate Studies & Research in Statistics,Mangalore University

4 2005 Sh. Mohd. Shahnawaz, Department of Statistics, ThirdLucknow University, Lucknow

5 2005 Sh. D. Satish, Acharya N. G. Ranga Agricultural ThirdUniversity College of Agriculture, Hyderabad

6 2006 Ms. Poonam Chandra, University of Madras, FirstChennai

7 2006 Ms. Nisha Malik, Ch. Charan Singh Agricultural SecondUniversity, Hissar, Haryana.

8 2006 Ms. Jyothi Bodani, Acharya Nagarjuna University, SecondNagarjuna Nagar, Andhra Pradesh

9 2007 Shri Anirvan Chakraborty, Indian Statistical Institute, FirstNew Delhi

10 2007 Ms. Ketaki Garg, Indian Statistical Institute, New Delhi Second

11 2007 Ms. Nisha Malik, Chaudhry Charan Singh Haryana SecondAgricultural University, Hisar

12 2007 Ms. Jagrity Das, Guwahati University Third

13 2007 Ms. Arshi Tahniyat Javed Iqbal Khan, Sant Gadge ThirdBaba Amrawati University, Amrawati

14 2007 Shri Sarvadanand Barnwal, Delhi University Third

15 2008 Ms. Tuhina Biswas, Indian Statistical Institute, First*New Delhi

16 2008 Ms. Sumita Bose, Indian Statistical Institute, New Delhi Second**

17 2008 Shri Rohit Patel, Indian Statistical Institute, New Delhi Second**

18 2008 Ms. Subrat Kumari Behera, Indian Agricultural Statistics Third**and Research Institute, New Delhi

*For writing an essay on “Impact of Industrial Development on Environment and Ecology and the Role of Statistics”

** For writing an essay on “Use of Statistics in Measurement of Economic and Social Development”

43

Statistics Day 2008

Statistics Day Function at Vigyan Bhawan, New Delhi

Statistics Day 2008

44

2005

45

Statistics Day 2008

Statistics Day 2008

46

Percentage of women aged 20-24

married by 18 years

47

Statistics Day 2008

Lakshadweep DES celebrating Statistics Day 2007