Humble beginning

What began as a small room in the Presidency College in the year 1931, now comprises buildings on several acres of land in four major cities of the country - Kolkata, New Delhi, Bangalore and Hyderabad. What began with a total annual expenditure of less than Rs. 250 in 1931, now has a total annual expenditure of over Rs. 15,000,000. What began with a solitary human computer working part-time in 1931, now comprises over 250 faculty members and over 1,000 supporting staff and several modern-day Personal Computers, workstations, mini computers, super-mini computers and mainframe computers. Impressive as these figures are, they convey little idea of the road traversed, the range of activities undertaken and the intimate relationship of the institute with the life of the nation. And to think that all this was the brainwave and vision of a single human being, the ‘solitary human computer’- the legendary physicist-turned-statistician, Prasanta Chandra Mahalanobis!

Prasanta Chandra Mahalanobis, born in a well-to-do progressive family in 1893, had gone to study in Cambridge in 1913 and in a couple of years’ time completed his Tripos in Physics. Just prior to his departure for India for a vacation, his tutor, WH Macaulay, drew his attention to some bound volumes of the journal Biometrika. Mahalanobis got so interested that he bought a complete set of the journal. These volumes were, subsequently, to play a decisive role in converting the physicist Mahalanobis to the statistician Mahalanobis.

Although Mahalanobis was due to return to Cambridge to pursue research in Physics, he did not go back. Instead, he accepted the post of a professor of Physics in the Presidency College. The first important work in Statistics in the modern sense was undertaken by Acharya BN Seal who sought for Mahalanobis’s help in the statistical analysis of examination results of the Calcutta University.

Establishment of ISI

Mahalanobis set up the Statistical Laboratory in the Presidency College sometime in the 1920s. On December 17, 1931, the Indian Statistical Institute was founded as a learned society and housed in the Statistical Laboratory. The institute was registered on April 28, 1932, as a non-profit-distributing learned society under the Societies Registration Act (XXI of 1860) and is now registered under the West Bengal Societies Registration Act (XXXVI of 1961), amended in 1964. Sir RN Mookerjee accepted the office of the president of the institute and held this office till his death in 1936.

Recognition of Statistics as a Key Discipline in India

During the 1920s, in fact till the middle of the 1930s, all or nearly all the statistical work done in India was done single-handedly by Mahalanobis. The early statistical studies included analyses of data on the stature of the Anglo-Indians, the meteorological data, the rainfall data, the data on soil conditions etc. Some of the findings of these early studies were of great impact on the control of floods, agricultural development et al, and led to the recognition of statistics as a key discipline.

ISI Begin to Flourish

Mahalanobis’s influence was so pervasive that the students of Physics began to take interest in Statistics, the most notable of them having been Subhendu Sekhar Bose. Later on, several talented young scholars including JM Sengupta, HC Sinha, RC Bose, SN Roy, KR Nair, K Kishen and CR Rao joined to form an active group of statisticians while Mahalanobis of course continued to be the nucleus. Theoretical research in Statistics began to flourish in the institute. Research on large scale sample surveys won Mahalanobis a Fellowship of the Royal Society. Design and analysis of agricultural experiments also bloomed and led to some international contacts, notably with Sir Ronald A Fisher.

In January 1938, the first Indian statistical conference was organised in Calcutta with Sir Ronald A Fisher as the president. Over the next five years, similar conferences were organised by the Indian Statistical Institute, which not only had their academic value but had immediate promotional value, too. At the Indian Science Congress in 1942, a separate section for Mathematics & Statistics was created with Mahalanobis as the sectional president. Three years later, a separate section was created for Statistics. Thus Statistics gradually won recognition in India and Mahalanobis as well as his ISI played a leading role-if not the leading role-in this evolution.

Even in the early days of the institute, Mahalanobis realised the importance of imparting training to the statisticians. During the years 1932-39, more than 150 government officials had come to Calcutta on deputation from all over India for short term training courses in statistical methods. The demand for advanced training in Statistics gradually increased. Through the efforts of the staff of the institute, a post-graduate department of Statistics was established in the Calcutta University in 1941. A Research and Training School was also set up in the ISI.

While, with the passage of time, the Government of India became more and more aware of the need for developing Statistics, it wanted the ISI to concentrate only on training and research in Statistics. But Mahalanobis viewed these roles as insufficient for the institute and wanted it to be involved in the projects of national interest, in which Statistics would have a key role to play. Mahalanobis’s view was that the ISI must be allowed to take up large-scale projects, which besides being socially useful would also help in research and training to be practical-oriented and not just bookish. This tussle lasted for nearly a decade, and Mahalanobis finally won it with the support of the then president of the ISI and a personal friend of his, CD Deshmukh.

The first exploratory sample survey in India was undertaken under the technical guidance of Mahalanobis as early as in 1937 for improving estimates of jute crop. This study won the praise of Sir RA Fisher who submitted a memorandum in 1938 to the Government of India supporting statistical studies in India with the words: “... everything depends on the future of the Statistical Institute”.

After the completion of the five-year jute survey in 1942, the Government of Bengal asked Mahalanobis to conduct a similar survey, which also included the estimation of the yield of paddy crop, in view of the great famine of Bengal in 1943. After much debate, sample surveys were ultimately accepted as the basis for official estimates of crop yield in West Bengal in 1948,and later in the other states of independent India.

Diversification of the scope of surveys and projects continued. At the request of the Government of Bengal in 1944, a survey of economic and social conditions in Bengal was undertaken to assess the impact of the severe famine of 1942-43. This survey yielded information of much social significance.

In 1945, an important project on population data began. The tabulation programme of the 1941 census had been severely curtailed as an economic measure. Fortunately, at the direction of MWM Yeates, the then census commissioner, a 2% sample of the individual census slips had been preserved, and the ISI was asked by the government to reconstruct the age and occupation tables on the basis of this sample. This involved the transfer of information to about seven million punched cards and then tabulating the results. As some sample slips had been incorrectly drawn and some had been missing, it became necessary to devise special methods of estimation. DB Lahiri provided leadership in this project, which was successfully completed in 1950.

Many other projects that yielded information of great social relevance pertained to rural indebtedness, road development, family budgets, traffic flow etc. At the request of the Government of Burma, the ISI helped in organising a sample survey of economic conditions in Rangoon in 1948-49.

The pioneering work on crop yield estimation done by the ISI attracted the attention of Prime Minister, Jawaharlal Nehru, who desired that comprehensive information relating to social, economic and demographic characteristics be collected by conducting sample surveys on a countrywide basis. The ISI was asked by the government to take charge of the design of these surveys and of processing of data thereof. The National Sample Survey (NSS) was established in 1950. This led to the regular conduct of the most comprehensive sample surveys in the world at that time, under the guidance of Mahalanobis, by NC Chakravarti, JM Sengupta, DB Lahiri, Nimai Ghosh, Mohanlal Ganguli, Ajit Dasgupta and S Raja Rao.

Works on theoretical statistics gained much momentum from the projects and surveys undertaken by the institute. Mahalanobis’s works on anthropometry led to the development of the D 2 statistic for measuring population affinities. This, in turn, culminated in the proof by RC Bose and SN Roy that the sampling distribution of the studentised D 2 is a non-central F-distribution. Crop surveys and agricultural experiments resulted in a great deal of theoretical research in multivariate analysis and construction of designs. Sample surveys also resulted in some seminal work of philosophical profoundness on randomness and methods of drawing representative samples. Mahalanobis himself pioneered these thoughts, and these are among his most lasting contributions to Statistics.

With the undertaking of large-scale projects and surveys, computational facilities also started improving. The ISI was the first institute in India to have acquired an electronic computer. The Hollerith electronic computer, HEC-2M, was in operation at the ISI since 1956. Acquisition of this electronic computer resulted in the development of numerical methods and facilitated data processing.

The ISI also pioneered the development of Statistical Quality Control (SQC) in India, that too as early as in 1935.

The role played by the ISI in the sphere of national planning has been no less commendable either. In 1940, Jawaharlal Nehru asked Mahalanobis to prepare a statistical commentary on the reports of the National Planning Committee. Nehru visited the ISI in 1946. Impressed with the activities and performance of the institute, he started to take a keen interest in it. In 1949, Nehru asked Mahalanobis to work as the honorary statistical advisor to his cabinet. This resulted in closer connection of the ISI withnational planning activities, and on March 17, 1955, Mahalanobis submitted to the union government the draft plan-frame, which was accepted as the basis for the formulation of the second Five-Year Plan of India.

The institute was declared as an institution of national importance by an Act of the parliament in 1959 and was vested with powers to hold such examinations and to confer such degrees and diplomas in Statistics as may be determined by the institute from time to time. Following this, the ISI, in collaboration with the International Statistical Institute, established an International Statistical Education Centre (ISEC) in Calcutta in 1950.

On August 16, 1960, the ISI initiated the courses in Statistics leading to B. Stat. and M. Stat. degrees. JBS Haldane, along with Mahalanobis, drafted the structure of these courses. The institute’s structure, philosophy and activities won international acclaim to the extent that when the first Institute of Statistics was set up in the United States of America by Gertrude Cox, the Indian Statistical Institute was used as its model. Perhaps this is the only occasion when a developing country has been taken as a model by a developed country.

By the end of the first three decades after its formation, not only was the ISI on solid ground, but Statistics also gained recognition as a key technology. The next three decades witnessed a tremendous diversification of the activities of the institute. In addition to the research work on sample surveys and design of experiments, research in multivariate analysis, statistical inference and probability gained tremendous momentum. Research work on Mathematics, especisally related to solving intricate technical problems in probability and statistics, also became a focal theme of the ISI. This resulted in pioneering developments in the area of linear algebra, particularly in generalised inverses of matrices.

Thus, we may well say, the ISI, which foresaw and established a symbiotic relationship between statistics and other branches of science, has over this long period of 75 years maintained just the same kind of relationship with the state and the society while all along the guiding inspiration has been the early visionary-Prasanta Chandra Mahalanobis.