Describes a new method of evaluation of scientific output by laboratories engaged in diverse fields of research. This method helps to evaluate those outputs which are quite recent and not amenable to citation analysis. For the purpose of analysis, impact factor of journals in which papers are published are considered. A method for normalisation of impact factor of journals has been described and, normalised impact factors have also been used for the purpose of analysis. It is found that in such analysis normalised impact factor tends to show better results compared to simple impact factor. The analysis helps us to generate numerous performance indicators such as average impact factor and normalised impact factor for each laboratory and the research complex such as CSIR as a whole; average impact factor and normalised impact factor for each scientist of a laboratory and the research complex; spectral distribution of papers falling within various ranges of impact factors and normalised impact factors. By comparing the performances over several years the trend of research activity of each laboratory can also be obtained.
Authors:K. Sen, Manju L. Agarwal and S. Chakraborty
In this paper, joint distributions of number of success runs of length k and number of failure runs of length k' are obtained by using combinatorial techniques including lattice path approach under Pólya-Eggenberger model. Some of its particular cases, for different values of the parameters, are derived. Sooner and later waiting time problems and joint distributions of number of success runs of various types until first occurrence of consecutive success runs of specified length under the model are obtained. The sooner and later waiting time problems for Bernoulli trials (see Ebneshahrashoob and Sobel ) and joint distributions of the type discussed by Uchiada and Aki  are shown as particular cases. Assuming Ln and Sn to be the lengths of longest and smallest success runs, respectively, in a sample of size n drawn by Pólya-Eggenberger sampling scheme, the joint distributions of Ln and Sn as well as distribution of M=max(Ln,Fn)n, where Fn is the length of longest failure run, are also obtained.